Warner et al v. Gross et al

Filing 648

FINDINGS OF FACT AND CONCLUSIONS OF LAW. The prerequisites of a successful lethal injection challenge under the Eighth Amendment have been made clear by the Supreme Court in the Baze, Glossip and Bucklew cases cited in t his order. The plaintiff inmates have fallen well short of clearing the bar set by the Supreme Court. Consequently, the Eighth Amendment, as construed and applied by the Supreme Court in its lethal injection cases, does not stand in the way of execution of the Oklahoma inmates in this case under Chart D of the Oklahoma lethal injection protocol. A final judgment will be entered forthwith. Signed by Judge Stephen P. Friot on 6/6/2022. (llg)

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 1 of 45

IN THE UNITED STATES DISTRICT COURT FOR THE
WESTERN DISTRICT OF OKLAHOMA
RICHARD GLOSSIP, et al.,
Plaintiffs,
-vsRANDY CHANDLER, et al.,
Defendants.

)
)
)
)
)
)
)
)
)

Case No. CIV-14-0665-F

FINDINGS OF FACT AND CONCLUSIONS OF LAW
Table of Contents
Page
I.

Introduction ........................................................................................................ 1

II.

Findings of Fact .................................................................................................. 1
A. Procedural history ....................................................................................... 1
B. The 2020 Protocol ...................................................................................... 3
C. The effect of vecuronium bromide as used in the protocol ........................ 4
D. The effect of potassium chloride as used in the protocol ........................... 5
E. The pharmacology and clinical uses of midazolam ................................... 5
F. The four recent Oklahoma executions........................................................ 7
1. The John Grant execution ................................................................... 9
2. The Bigler Stouffer execution .......................................................... 14
3. The Donald Grant execution............................................................. 17
4. The Gilbert Postelle execution.......................................................... 18
G. Glossip first prong: The challenge to midazolam as used per Chart D... 19
1. Dosage, speed of administration and duration ................................. 19
2. What is “induction”? What is “general anesthesia”? ...................... 20
3. Ceiling effect..................................................................................... 24
4. Pulmonary edema ............................................................................. 26
5. The reliability of midazolam as used per Chart D ............................ 30
i

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 2 of 45

H. The adequacy of the consciousness check ............................................... 35
I. The plaintiffs’ proposed alternative methods of execution ...................... 37
1. Fentanyl plus compounded or commercially procured
pentobarbital or sodium thiopental ................................................... 37
2. Firing squad ...................................................................................... 38
III. Conclusions of Law .......................................................................................... 40
IV. Conclusion ........................................................................................................ 42

ii

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 3 of 45

I.

Introduction
The plaintiffs, Oklahoma inmates awaiting execution, challenge the

constitutionality, under the Eighth Amendment, of Oklahoma’s protocol for
execution by lethal injection. Defendants are various Oklahoma officials charged
with responsibility for implementation of that protocol. After three trips to the
Supreme Court, much motion work, and a nonjury trial, the matter is now ready for
entry of final judgment. The court’s findings of fact and conclusions of law, entered
pursuant to Rule 52(a), Fed.R.Civ.P., are set forth below. The final judgment will
also be entered today.
II.

Findings of Fact
A. Procedural history
This action was filed in 2014, as a challenge to Oklahoma’s then-existing

protocol for execution by lethal injection. A preliminary injunction hearing was held
in December, 2014, resulting in denial of a preliminary injunction to block the
executions of Charles Warner and three other inmates. Denial of the preliminary
injunction was affirmed by the Court of Appeals. Warner v. Gross, 776 F.3d 721
(10th Cir. 2015). The Supreme Court affirmed that judgment. Glossip v. Gross, 576
U.S. 863 (2015). In the meantime, Charles Warner, having been denied emergency
relief, was executed by lethal injection in January, 2015.1
A series of mishaps, some more serious than others, with respect to
Oklahoma’s implementation of its lethal injection protocol resulted in a state grand
jury investigation and a protracted period of development of a new protocol. A new
protocol was promulgated in February, 2020. That precipitated the filing of the third

1

Other orders in this case are published as Glossip v. Chandler, 554 F.Supp.3d 1176 (W.D. Okla.
2021) (summary judgment); Glossip v. Chandler, 2021 WL 4760383 (W.D. Okla. Oct. 12, 2021)
(denying reconsideration, and other matters); and Glossip v. Chandler, 2022 WL 136886 (W.D.
Okla. Jan 14, 2022) (denying preliminary injunction).

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 4 of 45

amended complaint, doc. no. 325, which remains the operative complaint in this
case.
The nature and status of the claims asserted in the third amended complaint is
shown in this table:
Count I

Fifth2 Amendment Due Process claim based on asserted failure to disclose
sufficient information re: development of the protocol and execution procedures.
Dismissed per order at doc. no. 349.

Count II

Eighth Amendment claim asserting that constitutionally impermissible pain and
suffering will result from the use of the three-drug lethal injection protocol
(midazolam, vecuronium bromide and potassium chloride). Defendants’ motion
for summary judgment denied; Count II was tried to the court in February and
March, 2022.

Count III

Eighth and Fifth3 Amendment claim asserting “deliberate indifference” to the
serious medical needs of the plaintiffs. Dismissed per order at doc. no. 349.

Count IV

First, Fifth4 and Sixth Amendment claim asserting unconstitutional denial of
access to counsel and the courts. Summary judgment granted on motion of the
defendants per order at doc. no. 449.

Count V

18 U.S.C. § 3599 claim asserting intentional deprivation of right to counsel.
Summary judgment granted on motion of the defendants per order at doc. no.
449.

Count VI

Ex Post Facto claim under U.S. and Oklahoma Constitutions, based on
substitution of midazolam. Summary judgment granted on motion of the
defendants per order at doc. no. 449.

Count VII

Fourteenth Amendment Due Process claim based on use of midazolam instead of
barbiturate. Summary judgment granted on motion of the defendants per order
at doc. no. 449.

Count VIII

Religious freedom claim asserting violation of plaintiffs’ sincerely-held religious
beliefs resulting from necessity of proposing a feasible alternative method of
execution. Dismissed per order at doc. no. 349.

Count IX

Eighth and Fourteenth Amendment claim asserting that plaintiffs will be subjected
to constitutionally impermissible human experimentation. Summary judgment
granted on motion of the defendants per order at doc. no. 449.

2

As was noted at an earlier stage of this action, the Fifth Amendment does not apply to the
defendants in this case. Doc. no. 349, at 4 (Count I construed as asserted under the Fourteenth
Amendment).
3

See note 2. Plaintiffs elected not to persist with their Fifth Amendment Due Process claim. Doc.
no. 349, at 8.
4

See note 2.

2

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 5 of 45

Count X

First and Fourteenth Amendment claim asserting denial of right of access to
governmental information. Summary judgment granted on motion of the
defendants per order at doc. no. 449.

As can be seen, the only claim remaining for trial was the Count II claim,
asserting that execution by lethal injection, using the three-drug protocol specified
in Chart D of the 2020 protocol, violated the Eighth Amendment’s prohibition of
cruel and unusual punishment. See, doc. no. 629 (Final Pretrial Report), at 4, ¶ 5.
That claim was tried in a nonjury trial which began on February 28, 2022 and was
completed on March 7, 2022. Twenty-eight plaintiffs, all awaiting execution,
remained in the case at the time of trial.5
The trial transcript has been filed and the parties have submitted their post-trial
filings.6 Doc. nos. 643 and 644. The matter is ripe for decision.
B. The 2020 Protocol
The 2020 protocol, entitled Execution of Inmates Sentenced to Death, was
adopted with an effective date of February 20, 2020.7 PX 45.8 The protocol includes
5

The plaintiffs remaining in the case are: James A. Coddington, Benjamin R. Cole, Carlos CuestaRodriguez, Richard S. Fairchild, Wendell A. Grissom, Marlon D. Harmon, Raymond E. Johnson,
Emmanuel A. Littlejohn, James D. Pavatt, Kendrick A. Simpson, Kevin R. Underwood, Brenda
E. Andrew, Richard E. Glossip, Phillip D. Hancock, Alfred B. Mitchell, Tremane Wood, Wade
Lay, Ronson Kyle Bush, Scott Eizember, John F. Hanson, Mica Alexander Martinez, Ricky Ray
Malone, Clarance Goode, Anthony Sanchez, Michael Dewayne Smith, James Ryder, Richard
Rojem and Jemaine Cannon. Three plaintiffs were executed after the court denied the defendants’
motion for summary judgment but before this case was tried (John Grant in October, 2021, Donald
Grant in January, 2022 and Gilbert Postelle in February, 2022).
6

The trial transcript was filed on April 8, 2022, doc. no. 640. In these findings, references to the
trial transcript are: Tr. __, with a parenthetical reference to the name of the witness where
appropriate.

7

Within the Oklahoma Department of Corrections, the document as a whole is called a “policy,”
and the only part of the document the DOC refers to as a “protocol” is Attachment D, which
governs preparation and administration of the lethal injection drugs. Tr. 1042. In this litigation,
however, the entire document has been referred to generically as the protocol, so the court will
stay with that terminology.
8

In these findings, PX refers to plaintiffs’ exhibits and DX refers to defendants’ exhibits.

3

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 6 of 45

detailed provisions with respect to (i) the staffing required to conduct an execution,
(ii) qualifications of the Intravenous Team (“IV Team”), (iii) training of team
members, (iv) designation and escort of execution witnesses, (v) procedures to be
followed in the days and weeks preceding an execution, (vi) news media access, (vii)
IV insertion, (viii) procedure in the execution chamber, (ix) post execution
procedures, (x) after action reviews, and (xi) the preparation and administration of
the lethal injection drugs.
The lethal injection drug combination relevant here is that set forth in Chart
D, PX 45, Attachment D, p. 3. Chart D calls for the administration of the following
substances in the order shown:
1.
2.
3.
4.
5.

500 milligrams midazolam (via two syringes)
60 ml. heparin/saline
100 milligrams vecuronium bromide9 (via two syringes)
60 ml. heparin/saline
240 milliequivalents potassium chloride (via two syringes)
6. 60 ml. heparin/saline

C. The effect of vecuronium bromide as used in the protocol
Vecuronium bromide is a drug used, clinically, as a muscle relaxant. A typical
clinical dose is 20 to 30 milligrams. When administered in the dosage called for in
Chart D (more than ten times the dosage specified on the FDA-approved label),
vecuronium bromide functions as a paralytic, rendering the inmate unable to move
(and thus unable to give any indication of pain or other distress). As a paralytic,
vecuronium bromide will stop breathing (by stopping the diaphragm), but it will not
stop the heart. An inmate who is alive, sensate and paralyzed by vecuronium
bromide would have a sense of suffocation. (This is the “substantial, constitutionally

9

Chart D provides for the use of pancuronium bromide or rocuronium bromide as alternatives to
vecuronium bromide.

4

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 7 of 45

unacceptable risk of suffocation” noted by the Supreme Court in Baze v. Rees, 553
U.S. 35, 53 (2008).)
D. The effect of potassium chloride as used in the protocol
Potassium chloride is a salt which is used clinically to replace potassium if the
patient’s potassium level is too low. It can also be used in heart surgery to make the
patient’s heart stop for purposes of the surgery. Potassium chloride is used in the
Oklahoma protocol (at a dosage several times higher, and more concentrated, than
any clinical dose) to induce cardiac arrest–fatal heart stoppage if the inmate is still
alive when the potassium chloride is pushed. Administration of 240 milliequivalents
of potassium chloride to a sensate inmate will cause pain and suffering.
The administration of potassium chloride, causing the heart to stop, renders
vecuronium bromide essentially superfluous to the accomplishment of the purpose
of the execution. As Dr. Mark Edgar explained, the paralytic effect of vecuronium
bromide stops breathing, which is unnecessary if the inmate’s heart is going to be
stopped with potassium chloride. Tr. 208.
E. The pharmacology and clinical uses of midazolam
Gamma aminobutyric acid–GABA–is a neurotransmitter found in humans.
Specifically, GABA is an inhibitory neurotransmitter. It is released at synapses
between the neurons and “modifies electrical activity in the brain.” PX 64.10 It does
that by acting at specific receptor sites. As a benzodiazepine, midazolam is a potent
central nervous system depressant which exerts its action through GABAA receptors
in the central nervous system by chemically binding to the GABAA receptors. Those
receptors are “the key targets that mediate most of the clinically important effects of

10

Bai, et al., “Distinct Functional and Pharmacological Properties of Tonic and Quantal Inhibitory
Postsynaptic Currents Mediated by γ-Aminobutyric AcidA Receptors in Hippocampal Neurons,”
Molecular Pharmacology, vol. 59, 814 (2001).

5

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 8 of 45

IV anesthetics.” PX 474.11 The net effect of all this is that the flow of chloride ions
into the nerve cells is increased, inhibiting the neuron–the neuron “does not fire as
easily.” Tr. 21 (Dr. Stevens). Central nervous system activity is thereby depressed.
As described in the FDA-approved label (which the court finds to be a reliable
source of information about midazolam), midazolam is a Schedule IV drug which
is, as noted above, in the benzodiazepine category. DX 44. (Other benzodiazepines
include Valium and Xanax.)

When midazolam is “given intravenous as an

anesthetic induction agent, induction of anesthesia occurs in approximately 1.5
minutes when narcotic premedication has been administered and in 2 to 2.5 minutes
without narcotic premedication or other sedative premedication.” Id. at 4 (emphasis
added).
The FDA label states that midazolam may be used “either alone or in
combination with other CNS [central nervous system] depressants” in such
procedures as bronchoscopy, cardiac catheterization and suture of lacerations. Id. at
8. The FDA label, after admonishing practitioners in several places about the
importance of administering midazolam slowly, cautions that “rapid intravenous
administration” of midazolam–which is exactly that the Oklahoma protocol
requires–”may result in respiratory depression, airway obstruction and/or arrest.”
DX 44 at 18.12

Reactions to midazolam can include agitation, hyperactivity,

involuntary movements, “retching” and, in 2.6 percent of patients, “vomiting.” Id.
at 9, 16, 17. (The difference, if any, between “retching” and “vomiting,” as those
terms are used in the FDA label, is unclear.) A clinical dose of midazolam (much

11

Vyuk, et al., “Intravenous Anesthetics,” Ch. 23 in Miller’s Anesthesia (Elsevier Health Sciences,
2014), 653.
12

This statement by the FDA is supported by the literature. E.g., PX 396: Reves, et al.,
“Midazolam: Pharmacology and Uses,” Anesthesiology, vol. 62, 310 at 312 (1985): “Midazolam
produces some respiratory depression.”

6

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 9 of 45

lower than the dose called for in Chart D) will typically begin to take effect in thirty
to sixty seconds or not much longer. Tr. 635 (Dr. Antognini).
The FDA warns–in a black box warning on the midazolam label–that
intravenously-administered midazolam can cause respiratory depression and
respiratory arrest. DX 44, p. 2.
At the trial, there was considerable debate as to just what “induction” of
anesthesia means, as that term is used in the FDA label and elsewhere. The gist of
plaintiffs’ argument, as discussed later in these findings, is that “induction” of
anesthesia really means nothing more than getting the sedation started, and cannot
be taken to connote arrival at a deep plane of anesthesia (which plaintiffs assert
requires the use of other agents, such as anesthetic gasses). For present purposes, it
is worth noting that the FDA label states that: “If needed to complete induction,
increments of approximately 25% of the patient’s initial dose may be used; induction
may instead be completed with inhalational anesthetics.” DX 44 at 21 (emphasis
added).
F. The four recent Oklahoma executions
In the weeks before this case was tried beginning on February 28, 2022, four
Oklahoma inmates, three of whom were plaintiffs in this case, were executed by
lethal injection under the Oklahoma protocol. Those were: John Grant (executed on
October 28, 2021), Bigler Stouffer (December 9, 2021), Donald Grant (January 27,
2022) and Gilbert Postelle (February 17, 2022). Various combinations of plaintiffs’
and defendants’ witnesses viewed all four of these executions.
In assessing the testimony from witnesses who viewed or participated in these
executions, it should be borne in mind that little, if anything, relevant to the issue of
infliction of pain and suffering can be inferred from anything the inmate did (or,
more to the point, did not do) after the vecuronium bromide was pushed. Plaintiffs
correctly assert, and defendants do not dispute, that after the vecuronium bromide
7

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 10 of 45

takes effect, a sensate inmate would be unable to describe or otherwise communicate
about any pain he may be experiencing. Consequently, the probative eyewitness
testimony bearing on the issue of what the inmate may have been feeling as the
execution took place is the testimony describing observations beginning with the
injection of the first syringe of midazolam and ending with the injection of the first
syringe of vecuronium bromide.
It should also be noted, preliminarily, that the court’s findings as to the four
recent executions are based on testimony given by both fact and expert witnesses.13
The expert witnesses’ descriptions of what they saw are, course, based on their
observations of objectively verifiable facts, as viewed, in some respects, through the
lens of their professional expertise. Consequently, it is appropriate here to comment
briefly on the battle of the experts in this case. Rarely, in any field of litigation, does
a court see and hear well-qualified expert witnesses giving expert testimony as
squarely–and emphatically–contradictory, on the issues at the heart of the matter, as
in this case. Some of the opposing experts in this case have squared off in other
midazolam challenges in other courts around the country, in what amounts to a
midazolam roadshow.14 On the whole, the results in those cases have not been
favorable to the inmates awaiting execution, but extended analysis of how some of
the experts in this case have fared in other cases is not necessary here. There is one

13

Preliminary injunction hearings were held in this case on October 25, 2021 (relating to John
Grant, Julius Jones, Donald Grant and Gilbert Postelle) and on January 10, 2022 (relating to
Donald Grant and Gilbert Postelle). At the trial on the merits, the parties agreed that the testimony
of fact witnesses at those hearings would be deemed incorporated into the trial record. Tr. 1021,
1119. The findings the court now makes as to the four recent executions are based in part on the
fact witness testimony from those hearings.

14

For instance, three of plaintiffs’ expert witnesses in this case (Drs. Stevens, Van Norman and
Williams) and two of defendants’ experts (Drs. Antognini and Buffington) testified in the
McGehee case in Arkansas. McGehee v. Hutchinson, 463 F.Supp.3d 870, 908 (E.D. Ark. 2020),
appeal docketed, No. 21-1965 (8th Cir. April 30, 2021, argued and submitted January 12, 2022).

8

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 11 of 45

fresh face in this case, and a credible one at that. That is Dr. Ervin Yen, a practicing
anesthesiologist who has administered midazolam and observed its effects thousands
of times.
1. The John Grant execution
John Grant was executed on October 28, 2021. PX 802. He was served a
meal at 5:13 p.m. on October 27, the day before he was executed. Grant was still
“drinking soda” and eating at 10:15 p.m. on October 27. Jan. 10, 2022 Tr., doc. no.
585, at 181 (“1/10 Tr.”). At 3:15 p.m. on October 28, just before he was taken to the
execution chamber, Mr. Grant was eating chips and drinking soda (Mr. Pibb Xtra).
The Grant execution began shortly after 4:00 p.m.

The curtain between the

execution chamber and the viewing rooms was raised at 4:08 p.m. PX 802, p. 4.
The first syringe of midazolam was pushed about one minute after the curtain
was raised. PX 802, p. 4. Dr. Yen (observing the execution at the request of the
defendants) noticed, at that point, “bubbles … [of] a different color” in the IV line.
Tr. 1129. Grant appeared to lose consciousness in less than a minute. Soon after
that, Grant’s chest and abdomen moved up and down in a “rocking boat” motion,
something Dr. Yen has never seen in a person who was conscious. Tr. 1110, 1130.
The probable cause of that was obstruction of Grant’s breathing due to blockage of
his throat by his tongue.15 Tr. 1130. This was a natural result of administration of
500 milligrams of midazolam. Tr. 1139. Dr. Yen saw no purposeful movements
after the rocking boat motion started. Tr. 1131.

15

The possibility of airway obstruction as a result of administration of midazolam is not a concept
new to this case. Dr. Van Norman, an expert witness retained by plaintiffs in this case and in the
Arkansas case, testified about midazolam-induced airway obstruction in the Arkansas case.
McGehee, 463 F.Supp.3d at 883.

9

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 12 of 45

Plaintiffs have placed considerable emphasis on the fact that John Grant
“vomited” after the midazolam was pushed.

Doc. no. 594, at 10, 11.

The

Department of Corrections log reflects that the midazolam was pushed at 4:09 p.m.
and “Inmate vomiting @ 1610 [4:10 p.m.].” PX 802, p. 4. There was a fair amount
of debate at the trial as to whether this was actually propulsive vomiting (the forceful
ejection of gastric contents) or a more passive flow of gastric contents from Grant’s
stomach as his muscles relaxed, due to the effects of midazolam, while he lay
strapped down in a supine position on the gurney.
As has been discussed, the FDA label discloses that “retching” or “vomiting”
can occur as a result of administration of midazolam in a clinical setting. It is also
fair to say, in general, that most any adverse effect a clinical dose of midazolam
might have, as administered slowly, per the FDA’s repeated cautionary advice, will
be aggravated when a massive dose is pushed essentially all at once. And before
addressing this issue, it is worth noting that (i) whether it was propulsive vomiting
or something more passive, Grant’s substantial intake of food and beverages in the
hours before he was executed would naturally have predisposed him to loss of gastric
contents–regardless of whether midazolam was a causal factor–while he lay on the
gurney, (ii) whether it was vomiting or something more passive, the fact and cause
of the loss of gastric contents is of no moment, in the context of the issues in this
case, if this all occurred after Grant had been rendered insensate by midazolam, (iii)
there would in any event remain a question as to whether the vomiting, alone or in
combination with aspiration of gastric contents in the few minutes before death,
would amount to the severe pain and suffering which the Supreme Court has said
must be proven in a method-of-execution challenge, and (iv) there would in any
event remain a question as to whether this episode, during the John Grant execution,
even if it entailed the requisite severity of pain and suffering, is something that is

10

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 13 of 45

“sure or very likely” (in the sense those words were used by the Supreme Court) to
occur in future executions under Chart D.
Dr. Van Norman, who did not witness the John Grant execution, opined, based
on photographs showing liquid residue on the floor a few feet away from the gurney,
that Grant’s loss of gastric contents was the result of “propulsion” of vomit, not
passive regurgitation. Tr. 504. PX 37, p. 190 (photograph).16 She assumed, without
explanation, that this episode of “active vomiting” indicates that Grant was
conscious. Tr. 504.
Eyewitness accounts, taken together with a series of photographs of the
gurney and the floor around the gurney, are at odds with Dr. Van Norman’s opinion.
Justin Farris, the Oklahoma Department of Corrections Director of Operations, was
in the execution chamber. On the subject of John Grant’s vomiting, Farris testified
that the vomit came out of Grant’s mouth, onto the pillow, and “was running from
the floor out that way.” Tr. 991. Dr. Yen, also an eyewitness, testified that the vomit
was visible on the floor (in the area pointed out by Dr. Van Norman, referring to DX
37, p. 190) because it “hit the ground and spread.” Tr. 1133. It did not propulsively
come out of Grant’s mouth. Tr. 1134. Consequently, Dr. Yen described this as
regurgitation, not vomiting. Tr. 1132. The regurgitation occurred because Grant
lost muscle tone as a result of unconsciousness and hypoxia. Tr. 1135. He was
under general anesthesia–unconscious–when he was regurgitating.
The photographic evidence supports the testimony of Mr. Farris and Dr. Yen.
DX 37, p. 190 (Bates 3478) does show residue out to the side of the gurney (on
Grant’s right side), as described by Dr. Van Norman. The flow of the liquid across

16

At trial, pages of some exhibits (mostly ones which were viewed electronically) were referred
to by their PDF pagination. An example is PX 37, p. 190. The PDF pagination, although
frequently used for reference purposes at trial, does not appear in the hard copy of the exhibit. In
this instance, the hard copy of PX 37, p. 190 is Bates stamped 03478.

11

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 14 of 45

the floor in that direction is shown by p. 244 (Bates 3532). The fact that there was
an abundant supply of vomitus is shown by p. 241 (3529), which depicts the residue
of what was clearly a substantial quantity of liquids and solids on the sheet and the
pillow. (Obviously, the liquid and solids which remained on the pillow next to
Grant’s head were not propulsively ejected away from the gurney.) There was a
large enough quantity that the liquid also flowed over to the left (Grant’s left) side
of the gurney, as shown by the photos at pp. 187, 199 and 206 (3475, 3487 and
3494). Dr. Van Norman’s opinion that Grant experienced propulsive vomiting is
undermined by the credible testimony of eye witnesses, which is supported by the
photographic evidence.
The court also rejects Dr. Van Norman’s speculation that Grant was conscious
during this episode, whether it was vomiting or passive regurgitation. As these
findings already suggest, and as is discussed in more detail below, the effect of the
massive dose of midazolam was to render Grant unconscious before he lost his
gastric contents. In fact, that physiological result of administering that massive dose
of midazolam to a man with a full stomach was unsurprising.17 Tr. 1135.
17

This makes it unnecessary to dwell on the question of whether the discomfort of vomiting (and
probably aspirating vomit) while conscious and strapped down on a gurney would satisfy the
Supreme Court’s standard for severe pain and suffering. The court has no reason to revisit its
finding on this issue, in the context of a motion for preliminary injunction, in January, 2022:
If, contrary to the court’s findings (which are compelled by persuasive evidence), Grant
did experience conscious pain and suffering as he was executed, the interval within which
he would have experienced pain and suffering would have been very short–less (probably
quite a bit less) than five minutes. To be sure, the “pain and suffering” would have been
more than negligible, mainly because, being strapped down in a supine position, Grant
would have had no way to stop or otherwise alleviate the combined effect of
simultaneously regurgitating and struggling to breathe. But the court concludes that pain
and suffering of that duration and magnitude, if it occurred at all, falls short of the severity
required, for Eighth Amendment purposes, by the Supreme Court’s lethal injection
trilogy.
Glossip v. Chandler, 2022 WL 136886, at *5 (W.D. Okla. Jan. 14, 2022).

12

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 15 of 45

A consciousness check was performed at approximately 4:15 p.m. PX 802,
p. 4. The consciousness check consisted of a sternum rub and raising Grant’s
eyelids. The sternum rub, performed by the doctor assisting in the execution
chamber, was described by Mr. Farris: “he takes his knuckles through the sternum
and rubs it really hard.” 1/10 Tr. 189.
The doctor concluded that Grant was unconscious, which he indicated (per
agreed upon procedure) with a knock on the door. The announcement of Grant’s
unconscious state was made at 4:15. PX 802, p. 5. The vecuronium bromide was
pushed immediately. Id. Grant did not move after that. The two syringes of
potassium chloride followed at 4:18 and 4:19. He was pronounced dead at 4:21 p.m.
Id.
The placement of the straps on Grant’s chest did not impair the doctor’s ability
to perform the sternum rub. As was noted by Dr. Van Norman, PX 37, p. 186 (3474)
shows the large chest straps placed in such a way that there would have been little
room left to perform a sternum rub. But that photo does not show how the straps
were actually arranged during the execution. As credibly explained by Mr. Farris,
during the execution, the point of intersection of the four straps was lower than is
shown in this photograph. Tr. 990. Notably, Farris’s explanation is consistent with
the placement of the chest straps on Gilbert Postelle, as shown in PX 837, p. 7 (Bates
OAG – 024457). On Postelle’s chest, the four straps converge at a centrally located
metal fitting, leaving several inches of Postelle’s sternum area uncovered by straps.
At about the time the announcement of unconsciousness was made, Grant was
breathing, although his “cheek fluttered,” his mouth opened and then his “chest
settled”. 1/10 Tr. 19. The last round of drugs was pushed at approximately the 8 or
8 1/2 minute mark. At approximately the 12 minute mark in the execution process,
Grant was pronounced dead. 1/10 Tr. 19.

13

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 16 of 45

2. The Bigler Stouffer execution
Bigler Stouffer, accompanied by a chaplain, was executed shortly after 10:00
a.m. on December 9, 2021. The curtain was opened at 10:00 a.m. The first syringe
of midazolam was pushed at 10:01; the second was pushed at 10:04. PX 42, p. 4.
Dr. Yen, viewing the execution from the witness room, concluded that Stouffer was
unconscious very soon–30 to 45 seconds–after the midazolam was pushed. Tr. 1146.
This observation, which the court finds to be credible, is consistent with the
Department of Corrections log, which notes that Stouffer was “Speaking to Chaplain
and then started snoring.” PX 42, p. 5. This consisted of “deep sleep. Snoring.” PX
42, p. 4. Stouffer was confirmed to be unconscious at 10:07. As recorded in the
Department of Corrections log, this consisted of “Verbal, Sternum rub, pinch right
Forearm.” Id. As described by Dr. Yen, the consciousness check consisted of the
doctor putting his hands on Stouffer’s chest, applying pressure, and shaking him.
Dr. Yen could see Stouffer’s body shaking. Tr. 1146. The vecuronium bromide was
pushed immediately after Stouffer was determined to be unconscious, followed by
the two syringes of potassium chloride at 10:11 and 10:12. He was pronounced dead
at 10:16 a.m. Id.
This is as good a place as any to address a discrepancy relating to drug
nomenclature. This occurred with respect to the Stouffer, Donald Grant and Postelle
executions and, it is fair to assume, with respect to the John Grant execution. The
Oklahoma protocol provides for the use of vecuronium bromide or pancuronium
bromide or rocuronium bromide as the paralytic agent.18 PX 45, p. 39, ¶ (C)4(b).

18

Rocuronium bromide is also specified, along with vecuronium bromide, in the Alabama threedrug lethal injection protocol. Saunders v. Hamm, 2022 WL 493693, at *12 (M.D. Ala. Feb. 17,
2022). Likewise for Florida, Ohio and Tennessee. See, Brant v. Reddish, 2019 WL 4600366, at
*5 (M.D. Fla. Sept. 23, 2019); Conway v. Shoop, 2020 WL 3403210, at *3 (S.D. Ohio June 19,
2020); and King v. Parker, 467 F. Supp. 3d 569, 571 (M.D. Tenn. 2020).

14

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 17 of 45

The paralytic designated for use in the four recent executions was vecuronium
bromide.
The protocol calls for the setting of primary and backup IV catheters. Id., p.
42, ¶ F(4). The two sets of syringes (one for the primary IV line and one for the
backup) are to be attached, in the proper drug sequence, to manifolds which in turn
are connected to the IV lines. PX 837, p. 9 (OAG – 024459), relating to the Gilbert
Postelle execution, shows this arrangement:

The protocol requires each syringe to be put “in its corresponding place in
the shadow board which is labeled with the name of the chemical, color, chemical
amount and the designated syringe number.” PX 45, p. 40, ¶ D(3). As the
photograph suggests, the “shadow board” functions more or less as a color-coded
visual template for correct placement of the drug syringes on the IV manifold. The
syringes for the paralytic were properly labeled: vecuronium bromide. But the
15

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 18 of 45

corresponding yellow labels affixed to the shadow board said “50 mg Rocuronium
Bromide,” as shown in PX 837, p. 12 (OAG – 024462):

This discrepancy in nomenclature was addressed with Mr. Farris at trial. He
testified that he personally verified the drugs used in the Donald Grant and Postelle
executions. Tr. 996, 1032. This is consistent with the DOC’s logs of those
executions. Those logs, which allow the record keeper to indicate which of the three
authorized paralytics was used, indicate that vecuronium bromide was used. DX
159; PX 813. The same is true for the John Grant and Stouffer executions. PX 802;
PX 42. But for Oklahoma’s previous problems with using the right drugs, this
discrepancy probably wouldn’t have been worthy of mention at the trial of this case.
The displeasure of Department of Corrections Director Scott Crow with this
discrepancy in the nomenclature on the label on the shadow board was evident and

16

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 19 of 45

is understandable. Tr. 1038. The court is satisfied that vecuronium bromide was
used in each of the four recent Oklahoma executions.19
3. The Donald Grant execution
Donald Grant was executed, with a chaplain present, shortly after 10:00 a.m.
on January 27, 2022. The curtain was opened at 10:00. The two syringes of
midazolam were pushed at 10:03 and 10:06. PX 810, p. 4. This execution was
witnessed by Dr. Joseph Antognini. By Dr. Antognini’s account, which the court
finds credible, Grant became “more subdued” at the two minute point and seemed
to fall asleep at 10:05. He gave no appearance of consciousness after that. Tr. 736.
Grant soon started a “rocking boat” motion, which Dr. Antognini took to be an
indication of complete or partial airway obstruction. Tr. 736-37. This would have
been due to either partial airway obstruction or total airway obstruction in
combination with diminished respiratory effort. Tr. 737. Grant was confirmed to
be unconscious at 10:08. The log records the consciousness check as “sternum rub;
verbal or painful stimulation.” PX 810, p. 4. As described by Dr. Antognini, the
consciousness check consisted of a sternum rub and speaking loudly: “Donald,
Donald” (heard through the glass, with the microphone off). The chest straps did
not cover Grant’s entire sternum.

Tr. 749-50.

The person performing the

consciousness check also squeezed Grant’s right arm and wrist. Tr. 739.
The two syringes of vecuronium bromide were pushed at 10:09 and 10:10.
Grant stopped breathing at 10:10 and the two syringes of potassium chloride were
pushed at 10:12 and 10:13. He was declared dead at 10:16 a.m. PX 810, p. 5.

19

It is worth noting that the Supreme Court has observed, in this case, that “all agree” that
vecuronium bromide and rocuronium bromide are “functionally equivalent for purposes of this
case.” Glossip, 576 U.S. at 873.

17

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 20 of 45

4. The Gilbert Postelle execution
Gilbert Postelle was executed on February 17, 2022. The curtain was opened
at 10:00 a.m. The syringes of midazolam were pushed at 10:01 and 10:03, after
which Postelle wiggled his hands and feet for a time. PX 813, p. 4; Tr. 573-74 (Dr.
Van Norman). The log records that at 10:03, Postelle was snoring and that his eyes
were partially open. PX 813, p. 4.
Postelle stopped moving about 2 ½ minutes after the execution began. Then,
a few seconds later, one or more of his fingers moved. Tr. 1151 (Dr. Yen). Dr. Yen
concluded that this was not purposeful movement and that Postelle was unconscious
at that point. Tr. 1151. Mr. Farris described it as a “twitch.” Tr. 1018.
Dr. Van Norman noted movement in some of Postelle’s fingers at or soon
after 10:09, which she described as “a conscious movement.” Tr. 594-95. The court
evaluates that observation in light of Dr. Yen’s testimony that finger movement
would not be a disturbing thing to see in a patient under general anesthesia. Tr. 1108.
(It is common for patients to move while under general anesthesia, which is one of
the reasons surgical patients are also given muscle relaxants.

Tr. 677 (Dr.

Antognini). The movement in patients to whom a muscle relaxant has not been
administered can include arms, legs and fingers. Tr. 1108 (Dr. Yen).)
The consciousness check was performed at 10:06, with a notation in the log
that the check consisted of “sternum rub; physical and verbal stimuli.” PX 813, p.
4; Tr. 1019 (Farris). As described by Dr. Van Norman, the person performing the
consciousness check did a sternum rub and shook Postelle. Tr. 577, 592, 607 (per
Dr. Van Norman, a “mild to moderate” sternum rub). But she opined that the
sternum rub “was not appropriately carried out.” Tr. 601. In her view, it should
have been done for thirty seconds or more. Tr. 608. As described by Dr. Yen, the
consciousness check consisted of the doctor, using both hands, putting his hands on
18

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 21 of 45

Postelle’s chest twice and shaking him vigorously. Tr. 1153. Dr. Yen disagreed
with Dr. Van Norman’s criticism of the adequacy of the consciousness check. Tr.
1155.
Postelle was declared unconscious at 10:06 and the two syringes of
vecuronium bromide were pushed at 10:07 and 10:08, followed by the potassium
chloride at 10:10 and 10:11. He was pronounced dead at 10:14 a.m. PX 813, p. 5.20
G. Glossip first prong: The challenge to midazolam as used per Chart D
As will be seen, the court’s findings on the ultimate issue of whether
midazolam can be relied upon to render an inmate insensate to pain culminate in a
finding that midazolam, administered as specified in the Oklahoma protocol, does,
in fact, reliably achieve the anesthetic desired effect. But several matters should be
addressed preliminarily.
1. Dosage, speed of administration and duration.
The first thing that needs to be borne in mind about midazolam as used for
lethal injection under the Oklahoma protocol is that the Chart D dose is a massive
dose–many times higher than any clinical dose–which is injected in an exceedingly

20

At the October 25, 2021 preliminary injunction hearing (see n. 13, above), the court heard
testimony from Spencer Hahn with respect to the Alabama executions of Willie Smith and Ron
Smith. At the conclusion of that hearing, the court made findings which addressed, among other
things, Mr. Hahn’s testimony. See, transcript of October 25, 2021 hearing, doc. no. 537, at
142-145. There is no need to repeat here the court’s findings as to the matters testified to by Mr.
Hahn. The court adheres to those findings. In particular, the evidence received at the trial on the
merits in this case gives the court no reason to revisit its October 25 findings as to the significance
of the movements described by Mr. Hahn. On January 10, 2022, at a hearing on a motion for
preliminary injunction filed by Donald Grant and Gilbert Postelle, the court heard testimony from
Julie Gardner and Meghan LeFrancois with respect to the execution of John Grant. In an order
filed on January 14, 2022, the court made its findings–for purposes of the motion then before the
court–as to the John Grant execution. See, Order of January 14, 2022 at doc. no. 587 (Glossip v.
Chandler, 2022 WL 136886, at *2-*5 (W.D. Okla. Jan. 14, 2022)). The findings the court now
makes as to the execution of John Grant take into account the testimony of Ms. Gardner and Ms.
LeFrancois at the January 10 hearing.

19

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 22 of 45

short period of time. The speed with which midazolam is pushed into the IV lines
when used for lethal injection far exceeds the rate repeatedly specified–with
accompanying cautionary language–in the FDA label, as discussed above. As for
the duration of the desired effect of an injection of 500 milligrams of midazolam, it
is necessary only to note at this point that plaintiffs’ focus on the question of whether
midazolam can be used to maintain anesthesia (it can) ignores the fact that in an
execution by lethal injection with the Chart D drugs, anesthesia or deep sedation is
only required for a few minutes, as indicated by the chronologies of the four recent
Oklahoma executions. If the inmate is rendered insensate by the time the second
and third drugs are pushed and accomplish their purpose, the question of how much
longer the 500 milligrams of midazolam might have rendered him insensate is
irrelevant.
2. What is “induction”? What is “general anesthesia”?
Induction. The fact that midazolam is approved (by the FDA and in the
literature) for “induction of general anesthesia” triggered a semantic dispute at trial.
The debate at trial as to just what “induction” of anesthesia means, as that term is
used in the literature on midazolam, boils down to this: plaintiffs’ experts testified
that midazolam cannot be relied upon to render an inmate insensate to pain during
execution under the Oklahoma protocol. Defendants’ experts testified to just the
opposite. Consequently, both sides have quite understandably sought corroboration
from credible, unbiased sources.

Literature–not the least of which is the

FDA-approved label–stating that midazolam is suitable for use in “induction of
general anesthesia” is abundant. DX 44, p. 8.21 As has been noted, the FDA label
21

See also, e.g.¸ PX 396: Reves, et al., “Midazolam: Pharmacology and Uses,” Anesthesiology,
vol. 62, 310 at 317 (1985): “Midazolam may be used intravenously for the induction of anesthesia
(table 2). Induction is accomplished when there is unresponsiveness to command and loss of the

20

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 23 of 45

also tells us that “[w]hen midazolam is given intravenous as an anesthetic induction
agent, induction of anesthesia occurs in approximately 1.5 minutes when narcotic
premedication has been administered and in 2 to 2.5 minutes without narcotic
premedication or other sedative premedication.” Id. at 4 (emphasis added). If
“induction” connotes nothing more than the first, preliminary leg of the journey to a
deep plane of anesthesia, then defendants don’t get the benefit of the corroboration
which might be derived from the objective literature when it states that midazolam
is usable for induction. This all led to the following colloquy between the court and
Dr. Antognini:
THE COURT: Let me ask a question. Counsel on both sides have made
admirable attempts to teach me, through witnesses, the difference
between “induction,” on one hand, and “general anesthesia,” the
maintenance phase of general anesthesia, on the other hand. And Dr.
Antognini, you may not be able to answer this, and if you can’t, then
tell me. But I gather, first of all, from your testimony, that once an
induction agent is used for the induction, first of all, then there is
typically, if you will, a handoff to some other agent to maintain general
anesthesia; is that basically correct?
THE WITNESS: Yes.
THE COURT: So let me put it in very basic laymen’s terms, then. If
you have, to your satisfaction, induced anesthesia with midazolam and
eyelash reflex.” The Reves study treats as unremarkable the proposition that midazolam induces
anesthesia in unpremedicated patients (e.g., patients for whom no other induction agent is used):
“The dose of midazolam required to induce anesthesia is higher in unpremedicated healthy patients
(up to 0.3 mg/kg) than in premedicated patients.” Id. A midazolam dosage of 0.3 milligrams per
kilogram translates to a 30 milligram dose in a man weighing 220 pounds–less than one tenth the
dosage called for by Chart D.
And see, PX 344: Miyake, et al., “Electroencephalographic response following midazolaminduced general anesthesia: relationship to plasma and effect-site midazolam concentrations,” J.
of Anesthesiology, vol. 24, 386 (2009): “Midazolam is used for sedation and general anesthesia.”
This is an appropriate place to note that a few minutes spent reviewing the parties’ respective
critiques of the opposing experts (doc. nos. 643 and 644) will reveal that the scholarly studies cited
by the parties are generally susceptible to cherry picking–each side can find something to like in
most any of the studies.

21

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 24 of 45

are ready to but not yet administering the handoff agent, if you will, and
the doctor turns to you and says, “is it okay to make the incision in this
person’s abdomen”? What’s your answer?
THE WITNESS: I would say that if I could -- I’m going to answer your
question, but I want to qualify it, if I may.
THE COURT: Sure.
THE WITNESS: If it was a short -- let’s say that it was an incision in
the abdomen to drain an abscess, where it was going to literally take
30 seconds, then I might say “go ahead.” If it’s an incision for a major
abdominal surgery, where they’re going to be there for two hours, then
I would start the maintenance phase, because I know we’re going to be
there for a long time. So for a very short procedure, I could rely only
on the drug for that.
THE COURT: Well, then, am I to understand that a successful
induction as -- in that framework, then, a successful induction gets the
patient to the plane of anesthesia that you then want to simply
maintain?
THE WITNESS: That is correct.
Tr. 647-49 (emphasis added).
Dr. Antognini’s account is consistent with the meaning of “induction” as that
term is used in the medical literature.

For instance, a study comparing the

performance of midazolam with another benzodiazepine used the following
definition of induction: “Induction of anaesthesia was defined as complete with loss
of lid reflex and failure to respond to oral commands.” PX 395 (emphasis in
original).22 “Induction” does not connote just the first step on the journey to a deep
plane of anesthesia.
General anesthesia. The American Society of Anesthesiologists (ASA) has
published a chart titled “Continuum of Depth of Sedation: Definition of General

22

Reves, et al., “Comparison of Two Benzodiazepines for Anaesthesia Induction: Midazolam and
Diazepam,” Canadian Anaesthetists’ Society Journal, vol. 25, no. 3 (May, 1978), at 211.

22

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 25 of 45

Anesthesia and Levels of Sedation/Analgesia.” That chart, PX 496, defines general
anesthesia as follows:
General Anesthesia is a drug-induced loss of consciousness during
which patients are not arousable, even by painful stimulation. The
ability to independently maintain ventilatory function is often impaired.
Patients often require assistance in maintaining a patent airway, and
positive pressure ventilation may be required because of depressed
spontaneous ventilation or drug-induced depression of neuromuscular
function. Cardiovascular function may be impaired.
The ASA chart, PX 496, summarizes the gradations of sedation and anesthesia
in terms of patient responses and potential complications as follows:

With the benefit of the ASA chart, the abundant professional literature in the
record, and the testimony of several of the experts who testified at trial, the court
finds, quite readily, that (i) the ASA’s definition of general anesthesia comports with
the sense in which that term was used at trial, and (ii) if an inmate being executed
per Chart D of the Oklahoma protocol reaches a state of general anesthesia as defined
by the ASA, he will not be sensate to pain resulting from the injection of any of the
Chart D drugs (or, for that matter, pain from any other source).

23

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 26 of 45

3. Ceiling effect
Seven years ago, in this case, the Supreme Court noted that “Petitioners assert
that midazolam’s ‘ceiling effect’ undermines the District Court’s finding about the
effectiveness of the huge dose administered in the Oklahoma protocol. Petitioners
argue that midazolam has a ‘ceiling’ above which any increase in dosage produces
no effect.” Glossip, 576 U.S. at 887. Justice Alito’s framing of the “ceiling effect”
issue is as important now as it was then: “The relevant question here is whether
midazolam’s ceiling effect occurs below the level of a 500-milligram dose and at a
point at which the drug does not have the effect of rendering a person insensate to
pain caused by the second and third drugs.” Id.
Plaintiffs’ contentions about the ceiling effect deserve attention because those
contentions tie directly into their argument that midazolam “is unable to render a
person insensate to pain no matter how much of it is injected into a person.” Doc.
no. 594, at 13.
As an initial matter, it is unremarkable that, for midazolam or any other drug,
there would be a dosage above which an incremental increase in the amount
administered would fail to produce a commensurate increase in the effect of the drug.
Consequently, there is no reason to dwell on the question of whether midazolam has
a linear dose response. Another beginning point for this analysis is the fact that the
defendants make no bones about the fact that the 500 milligrams specified in Chart
D is a massive dose–far above any clinical dose ever administered and likely well
above the dosage at which the incremental effects of increases in dosage taper off to
nil. As Dr. Michael Weinberger, one of plaintiffs’ experts, explained: “[A]bove a
certain dose, you would not see a clinical effect.” Tr. 399. Although the evidence
is in conflict as to the level at which the ceiling effect sets in, there is no reason to
question that fundamental proposition. On that score, Dr. Weinberger referred to a
study of the effect of midazolam when used in combination with halothane (an
24

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 27 of 45

anesthetic gas) in hysterectomy patients. The authors found that “[f]urther increases
in midazolam concentration continued to reduce the MAC [minimum alveolar
anesthetic concentration] of halothane but to a lesser degree.” PX 270 (emphasis
added).23 The short of the matter is that, once the GABAA receptors are saturated,
additional midazolam will not chemically bind to the GABAA receptors and
consequently will have no incremental effect as a central nervous system depressant.
Consistent with Dr. Weinberger’s testimony, Dr. Antognini, called by the
defendants, did not dispute that there are data showing a plateauing or a ceiling
effect. Tr. 671. But, as Dr. Antognini explained: “[T]he question is can you achieve
deep -- or unconsciousness, deep levels of anesthesia, whatever term you want to
use, to a point where a person would not have the conscious awareness of pain from
a noxious stimulus, including the potassium chloride that is used in the Oklahoma
protocol.” Tr. 671. Dr. Antognini’s elaboration is on all fours with Justice Alito’s
treatment of the ceiling effect seven years ago, in this case, as quoted above.
Plaintiffs’ contentions as to midazolam’s ceiling effect are, thus, of no
moment unless the ceiling effect sets in while the inmate can sense the pain he would
endure as a result of the administration of the massive doses of vecuronium bromide
and potassium chloride. The evidence satisfies the court that midazolam’s ceiling is
higher than the dosage required for midazolam to have the effect intended in the
Oklahoma protocol.
The most persuasive evidence that a ceiling effect is not problematic in
executions under Chart D is the evidence establishing to the satisfaction of the court
that midazolam does reliably place inmates under general anesthesia before the
vecuronium bromide and potassium chloride are pushed, as discussed in part (5),

23

Inagaki, et al., “Anesthetic Interaction Between Midazolam and Halothane in Humans,”
Anesthesia and Analgesia, vol. 76, 613 at 615 (1993).

25

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 28 of 45

below. Once anesthesia is induced as a result of the administration of midazolam,
the injection of additional midazolam will not diminish the effect the drug has
already had. Tr. 882 (Dr. Buffington). As Dr. Yen put it, in the context of
vecuronium bromide: “You know, if I gave you 20 milligrams of vecuronium, you
would not breathe, you would be paralyzed. If I gave you 50 milligrams of
vecuronium, a muscle relaxant that paralyzes you, you would not be more paralyzed.
You wouldn’t be less paralyzed.” Tr. 1094.
As measured by the bispectral index (BIS), a person is under general
anesthesia at a BIS level of 60 to 40.24 Tr. 35 (Dr. Stevens), Tr. 756, 757 (Dr.
Antognini), Tr. 891 (Dr. Buffington). Midazolam has been demonstrated to be
capable of inducing a BIS score well below 60 and in some cases below 40. Tr. 892894 (Dr. Buffington). Dr. Yen has never used midazolam to get a patient to a BIS
level lower than 70, and he wouldn’t want to do that, because “there’s a chance that
the patient could stop breathing, be obstructed, so that’s not necessarily my
endpoint.” Tr. 1243-44. In the case at bar, we are not, of course, concerned with
the possible adverse health effects of a midazolam dosage high enough to get the
inmate’s BIS level below 60. The relevant fact is that, regardless of what BIS level
might be achieved with a clinical dose of midazolam (and that will vary from patient
to patient), the 500 milligram dose specified in Chart D is, by design, more than
sufficient to bring about all of the suppression of central nervous system activity that
midazolam is capable of achieving.
4. Pulmonary edema
Pulmonary edema is a condition in which there is “excess fluid in the lungs.”
Tr. 197 (Dr. Edgar). The physiological effect of pulmonary edema is that “there’s a

24

If a person is fully awake, his BIS level is 100. If he has no brain activity at all, his BIS level is
zero. Tr. 32. BIS is determined on the basis of electroencephalographic input. Tr. 771.

26

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 29 of 45

problem moving oxygen into the bloodstream and carbon dioxide out of the
bloodstream.” Tr. 1157 (Dr. Yen). Pulmonary edema can be detected on autopsy:
“[W]hen you section [lungs] with a knife, you may find fluid flows out if it’s -- if
they’re quite edematous. And if they’re somewhat less edematous, you give them a
squeeze with your hand, then the fluid comes out.” Tr. 201 (Dr. Edgar).
Dr. Edgar’s review of the reports of the autopsies from thirty-two executions
conducted using midazolam disclosed that twenty-seven of the inmates had
pulmonary edema. Tr. 211. And he cannot rule out that the remaining five also had
pulmonary edema. Tr. 212. Dr. Edgar also personally performed the autopsy of an
Ohio inmate. He also had pulmonary edema. Tr. 196-97, 252. (Dr. Edgar sees
pulmonary edema in about 50 percent of all the autopsies he performs. Tr. 201.)
Dr. Edgar testified that pulmonary edema can be caused by heart failure
(“cardiogenic” pulmonary edema) or by “a chemical or a physical problem.” Tr.
197. As for the “chemical” problem, early on, Dr. Edgar posited that one possible
cause of pulmonary edema in an inmate undergoing execution could be the acidity
of midazolam. This was his “initial thought.” Tr. 270. But, in Dr. Edgar’s
estimation, with the benefit of the four recent Oklahoma executions, pulmonary
edema induced by the acidity of midazolam has been demoted to “one possibility.”
Id. He elaborated:
THE COURT: So coming back to your -- the evolution, if you will, of
your analysis, based on the four recent Oklahoma executions, I take it
you have a somewhat reduced level of confidence that the acidity of the
midazolam is a directly causal factor?
THE WITNESS: Exactly right. I think you might ask yourself, well,
gee, that seems like a dramatic change, why the sudden change? Well,
the only information available to me regarding the mechanism of
pulmonary edema in these cases is the executions, the observations
made during them and the anatomic observations made after them. So
each one is a potentially important data point.
27

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 30 of 45

Tr. 271-72.
Thus, as to acid-induced pulmonary edema, plaintiffs did contend at one point
that the acidity of midazolam caused edema in inmates undergoing execution with
midazolam. Third Amended Complaint, doc. no. 325, ¶ 69. But, in light of Dr.
Edgar’s testimony (quoted above), plaintiffs have now acknowledged that
“[e]vidence from the four recent executions in Oklahoma has reduced Dr. Edgar’s
confidence in his earlier theory that the pulmonary edema might be caused by an
acidic injury to the capillaries of the lungs when midazolam is injected at a high
volume and at a rapid pace.” Doc. no. 643, at 7-8. This is understandable, given the
fact that Dr. Edgar now goes no further than to offer that acid-induced pulmonary
edema is “one possibility.” Tr. 271. Plaintiffs thus concede that Dr. Edgar “cannot
explain the mechanism with certainty at this point.” Doc. no. 643, at 8. The only
remaining “mechanisms” posited by Dr. Edgar (and to which he testified with
noticeably greater certainty) are cardiogenesis and negative pressure in the lungs,
caused by obstructed breathing.25 Tr. 197-98, 200, 267-69. As discussed below, Dr.
Yen agrees.
Dr. Edgar very cogently explained that negative pressure pulmonary edema
brings fluid out of the small blood vessels and into the lungs “by the person trying
to breathe in against an obstruction. And when you do that, it creates a very high
negative pressure, almost a partial vacuum, if you will, in the airways and air spaces
in the lung. And that vacuum can suck fluid, if you will, out of the capillaries in the
lungs.” Tr. 268.

25

Even if the acid induction theory remained in contention, the court would readily conclude that
plaintiffs have not carried their burden of proof on this point, for the reasons credibly articulated
by Dr. Joseph Antognini (Tr. 694) and Dr. Daniel Buffington (Tr. 920-22).

28

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 31 of 45

Dr. Yen, called by defendants, agreed with Dr. Edgar that pulmonary edema
can be cardiogenic or can result from negative pressure. As for negative pressure
pulmonary edema, “[i]f a person obstructs their airway or partially obstructs, then
the diaphragm contracting is actually causing a vacuum in the chest and that can
actually suck some fluid out of the blood vessels and into the lung. And I’ve seen
that. And that takes minutes, not seconds.” Tr. 1157 (Dr. Yen). In inmates
undergoing execution, Dr. Yen believes that pulmonary edema is caused by negative
pressure or “cardiogenic shock from the heart not pumping adequately because of
lack of oxygen.” Tr. 1159. Consistent with that, Dr. Yen noted obstruction of John
Grant’s breathing, and Dr. Van Norman noted that Gilbert Postelle was “trying to
breathe against an obstructed airway.” Tr. 1130, 575.
Pulmonary edema can be mild or severe. Dr. Edgar acknowledged that mild
pulmonary edema does not entail severe pain and suffering. Tr. 286. In a clinical
setting, as pulmonary edema “becomes more severe clinically, it -- the patients begin
to feel a sense of -- a sense of doom, a sense of panic, terror, drowning, asphyxiation
on the severe end. Those are manifestations of severe pulmonary edema.” Tr. 198.
Importantly, Dr. Edgar could not say when in the lethal injection process
(prior to injection of the vecuronium bromide) the pulmonary edema began to
develop in the inmates whose autopsy reports he reviewed. Tr. 316-17. To his
credit–because he did not purport to be an expert on the effects of midazolam–Dr.
Edgar could not say whether pulmonary edema had set in “before any given inmate
became unconscious or unaware of pain.” Tr. 317. He ventured no guess as to what
the sensations of pulmonary edema would be in person sedated with a central
nervous system depressant. Tr. 289. Dr. Yen testified (credibly, as discussed below)
that “those two ways [cardiogenic and negative pressure] of getting pulmonary

29

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 32 of 45

edema take minutes, whereas the onset of the unconsciousness is in seconds.” Tr.
1159-60.
On the basis of all of the expert testimony addressing the point, it is safe to
say that pulmonary edema is found in many inmates–it might well be the vast
majority–executed by lethal injection. Cardiogenic pulmonary edema results from
impairment or cessation of cardiac activity. Tr. 197-98, 269, 284 (Dr. Edgar); 1157
(Dr. Yen). Negative pressure pulmonary edema results from obstruction of the
airway in an unconscious inmate whose tongue–because he is unconscious and
supine–has settled into the back of his throat. Dr. Antognini persuasively explained
that a patient who is conscious can keep his own airway open. Tr. 681-82. This is
why Dr. Yen has never seen the “rocking boat” motion (indicative of an obstructed
airway) in a conscious patient. Tr. 1110. In contrast, if a person is unconscious, as
Dr. Edgar put it, “their tongue can fall back,” with the result that “[t]hey’re less able
to keep their airway patent.” Tr. 310-11. The court is persuaded that pulmonary
edema from either cause–cardiogenesis or negative pressure–happens (if it happens
at all) when the inmate is insensate to pain. Tr. 1159-60 (Dr. Yen).
5. The reliability of midazolam as used per Chart D
The ultimate question before the court on Glossip’s first prong is whether
midazolam, used as per the Oklahoma protocol, reliably renders the inmate insensate
to pain before the vecuronium bromide is pushed and before pulmonary edema sets
in, if it sets in at all. We know that Oklahoma’s three-drug Chart D protocol does
cause death and that–absent a Lockett-type26 IV access mishap–it causes death
within a few minutes after the midazolam is pushed. That leaves the question of

26

The execution of Clayton Lockett, though ultimately successful, encountered exceptional
difficulties in establishing IV access, with the result that Lockett was not pronounced dead until
about 45 minutes after the midazolam was pushed. See, doc. no. 179, at 10-17.

30

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 33 of 45

whether the inmate is exposed to a constitutionally unacceptable risk of a
constitutionally unacceptable level of pain during that interval. If the inmate is
insensate to pain, it does not matter, under the Fourth Amendment, that he may be
paralyzed, that his lungs may be filling with fluid, that he has been injected with 240
milliequivalents of potassium chloride, or that the injection of the last few hundred
milligrams of midazolam had no incremental effect.
The opposing experts who testified at trial had respectable paper credentials,
although Dr. Stevens brought to the courtroom substantial grist for cross
examination.27 Among the experts called by plaintiffs, Dr. Gail Van Norman was
the most emphatic in opining that “to a virtual medical certainty, all four of [the
recently-executed Oklahoma inmates] experienced extreme pain and suffering in the
form of sensations of asphyxiation, drowning, the onset of paralysis in the injection
of potassium chloride during their execution and that they were awake and conscious
but unable to respond during that period of time except for the movements that we
saw on the part of John Grant and Gilbert Postelle.”28 Tr. 624. The court is
unpersuaded. The evidence persuades the court, and not by a small margin, that
even though midazolam is not the drug of choice for maintaining prolonged deep
anesthesia, it can be relied upon, as used in the Oklahoma execution protocol, to
render the inmate insensate to pain for the few minutes required to complete the
execution.

27

See, e.g., Tr. 112-13 (Dr. Stevens describing some of his own midazolam-related work as
“amateurish” and expressly disavowing his earlier work in a midazolam case), 114
(acknowledging lack of acceptance of his work relating to midazolam by courts in three states).
28

It is worthy of note, though determinative of nothing in the case at bar, that, in McGehee, “Dr.
Van Norman conceded that she has no direct scientific data to support the proposition that any
inmate experienced severe pain and suffering during an execution.” McGehee, 463 F.Supp.3d at
913.

31

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 34 of 45

Although Dr. Craig Stevens, called by the plaintiffs, acknowledged that
midazolam can produce deep sedation (Tr. 14, 47), plaintiffs’ witnesses, on the
whole, maintained that midazolam as used in the Oklahoma protocol cannot render
the inmate insensate to pain during the execution process. As to the pharmacological
effects of midazolam, plaintiffs’ main witnesses were Dr. Stevens and Dr. Michael
Weinberger. Dr. Stevens, a pharmacologist, relied mainly on published studies.29
Dr. Weinberger, who practices pain medicine, has not put a patient under anesthesia
since 1998. Tr. 362, 427. He has no recent experience with midazolam, nor has he
done research relating to midazolam, or, for that matter, any other benzodiazepine.
Tr. 427-28.
For her part, Dr. Van Norman sought to support her opinion, “to a virtual
medical certainty,” that Donald Grant experienced severe pain and suffering, by
citing the fact that “he continued to try to breathe all the way up until the time he
was paralyzed by the second drug of the protocol.” Tr. 567-68. Her reliance on the

29

At trial, the court took a permissive approach–applicable equally to both sides–to the admission
of various clinical studies and treatise segments into evidence. For the most part, if not in every
instance, the experts testified that they relied, in forming their opinions, on the studies and treatises
they cited in their reports or in their testimony. Preferring to evaluate the weight to be given to the
studies and treatises in light of the expert testimony and other evidence taken as a whole, the court
did not in every instance require a full Rule 703 incantation as a prerequisite to admission of the
studies and treatise excerpts. Cf. McGehee v. Hutchinson, 463 F.Supp.3d 870, 908 (E.D. Ark.
2020), appeal docketed, No. 21-1965 (8th Cir. April 30, 2021, argued and submitted January 12,
2022). It is worth noting that the studies relied upon by both sides can be distinguished from the
facts of this case in several ways. Perhaps most importantly, they can be distinguished in terms of
dosage, both in absolute terms and in terms of differences between doses administered to study
subjects. They can also be distinguished in terms of the study subjects. In some of the studies, the
subjects were animals, such as mice (PX 379), squirrel monkeys (PX 379), rats (PX 490) and dogs
(PX 428). At least one study used an in vitro technique (e.g., petri dish). PX 64. Some of the
studies can be distinguished because they involved midazolam in combination with other drugs by
way of premedication or otherwise. Some of the studies can also be distinguished because they
don’t involve midazolam at all. There is, obviously, no reason to expect that any of the experts’
research would have turned up a clinical study focusing on the effects of the injection of a 500
milligram bolus dose of midazolam into a human being.

32

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 35 of 45

fact that Donald Grant was breathing as an indication that he was conscious when
the vecuronium bromide was pushed struck the court as wholly untenable. She
backed off from that position when she was obliged to acknowledge that, in her
clinical practice, she sees insensate patients breathing all the time. Tr. 568.
As to the reliability of midazolam to serve its intended purpose under the
Oklahoma lethal injection protocol, the court finds, after careful consideration of all
of the expert testimony, that the testimony of Dr. Ervin Yen and Dr. Joseph
Antognini is the best-supported and most persuasive. The persuasive value of their
testimony is rooted in their extensive professional experience administering
midazolam, their observations of its effects on their own patients, and, not least, their
credible observations from the executions of John Grant (Dr. Yen), Bigler Stouffer
(Dr. Yen), Donald Grant (Dr. Antognini) and Gilbert Postelle (Dr. Yen).30
Dr. Yen estimates that he has administered anesthesia to more than 25,000
patients, using midazolam with “about a quarter of those patients.” Tr. 1081. Dr.
Yen has about fifteen to twenty patients a week and uses midazolam with about a
quarter of those patients in a normal week. Tr. 1081-82. For his part, Dr. Antognini
has served as a professor of anesthesiology at the University of California at Davis
and has published research “in the field of mechanisms of anesthesia. I was primarily
interested in where, of the central nervous system, where anesthetics work, in the
brain, in the brainstem, in the spinal cord, things of that nature.” Tr. 626.

30

Plaintiffs’ critique of Dr. Yen’s qualifications, doc. no. 643 at 62, is wholly unpersuasive. (The
implicit suggestion that Dr. Yen’s credibility is somehow diminished by his public service borders
on the offensive. Id.) The court’s confidence in Dr. Yen’s expertise (and his credibility, a slightly
different matter than expertise) is based predominantly on (i) his vast experience with midazolam
and other anesthetic agents and (ii) his cogent explanation of the basis for the opinions to which
he testified. The court cannot but note, also, that Dr. Yen was chosen by four of his colleagues in
the practice of anesthesiology to be their anesthesiologist for their heart surgeries. Tr. 1075-76.

33

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 36 of 45

Midazolam takes effect quickly, but, in doses sufficient to induce general
anesthesia, it has a long recovery time. Tr. 634-35, 653 (Dr. Antognini). Thus, its
main clinical use is as a sedative, because a dose sufficient to take the patient to a
state of general anesthesia would take too long to wear off. Tr. 641-43 (Dr.
Antognini). As Dr. Yen put it, “patients might stay in the recovery room for too long
or in the hospital for too long or at the outpatient surgery place for too long. And the
longer a patient stays in the hospital, the more expense that’s incurred. The longer
they stay in the recovery room, the more expense that’s incurred.” Tr. 1103.
Consequently, it is “pretty unusual” now to use midazolam as the sole drug to induce
anesthesia. Tr. 641 (Dr. Antognini).
The FDA label for midazolam gives separate treatment to midazolam’s
sedative effect and its use as an anesthetic induction agent. DX 44, p. 4. The FDA
label tells us that “induction of anesthesia” occurs in “2 to 2.5 minutes without
narcotic premedication or other sedative premedication.” Id.31 As is discussed
above, “induction” means arrival at a deep plane of anesthesia, not the first step on
that journey. See also, Tr. 1084 (Dr. Yen). Dr. Antognini and Dr. Yen concur with
the FDA’s statement. The matter was articulated well by Dr. Yen. As Dr. Yen put
it (and as the court has found), “I heard a lot of talk this week about induction and
maintenance. And what ‘induction’ to me means, is that before the induction, you’re
awake; after the induction, you are in a state of general anesthesia, according to the
chart that you’ve seen from the American Society of Anesthesiologists. And so I’ve
used it that way.” Tr. 1084. Using “induction” in that sense, Dr. Yen testified, and

31

In the framework of low clinical dosage, Dr. Yen agrees with this time estimate. “[I]t might
take a minute or two or even three.” Tr. 1104. But “the higher the dosage, the quicker it works.”
Id.

34

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 37 of 45

the court finds, that “midazolam alone can be used to induce general anesthesia.”
Tr. 1086.
Although deep sedation would suffice to eliminate pain from the lethal
injection procedure, 500 milligrams of midazolam will easily accomplish general
anesthesia. Tr. 1105, 1157 (Dr. Yen). The recipient of that dose will feel no pain.
Tr. 1105 (Dr. Yen). Vecuronium bromide will not raise an unconscious person to
consciousness, nor will it reverse the effects of midazolam. Tr. 1120 (Dr. Yen).
Likewise, injection of potassium chloride will not reverse the effects of midazolam
and will not cause an anesthetized person to feel pain. Tr. 1122-23 (Dr. Yen).
The credible evidence persuades the court that there is a high probability, with
respect to each of the four recent Oklahoma executions, that the Chart D sequence,
beginning with the injection of 500 milligrams of midazolam, worked as intended.
It is highly probable that the inmates felt no physical pain other than that associated
with the insertion of the IV lines. Tr. 744 (Dr. Antognini, re: Donald Grant), 114041, 1148, 1155 (Dr. Yen, re: John Grant, Stouffer and Postelle). Putting the matter
in terms of the high bar set by the Supreme Court in Baze and reiterated by the Court,
in this case, seven years ago, the evidence precludes any finding that execution by
lethal injection under Chart D of the Oklahoma protocol is sure or very likely to
cause needless suffering. Glossip, 576 U.S. at 877 (citing Baze, 553 U.S. at 50). It
is highly probable that an inmate undergoing lethal injection with the Chart D
combination of drugs will become insensate to pain within a very short time after
the midazolam is pushed and will remain in that condition until injection of the three
drugs causes death.
H. The adequacy of the consciousness check
Common sense tells us that a consciousness check is an important safeguard
against infliction of needless suffering during execution by lethal injection.
35

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 38 of 45

The Oklahoma protocol requires the IV Team leader to “physically confirm
the inmate is unconscious by using all necessary and medically-appropriate
methods.” PX 45, Attachment D, p. 7. The IV Team must consist of individuals
from any of six medically-related fields specified in the protocol: physicians,
physician assistants, nurses, emergency medical technicians, paramedics and
“Military corpsman or other certified or licensed personnel including those trained
in the United States military.” PX 45, p. 6. All members of the IV Team “shall be
currently certified or licensed within the United States.”

Team members are

screened on the basis of “a review of the proposed team member’s qualifications,
training, experience, and/or any professional license(s) and certification(s) they may
hold.” Id.
Plaintiffs complain that the protocol does not require the IV Team leader to
“have a background in anesthesiology or pain management.” Doc. no. 643, at 76.
As a baseline, the court will note that the consciousness check which passed muster
in Baze consisted of a “visual inspection performed by the warden and deputy
warden.” Baze, 553 U.S. at 59. And Chief Justice Roberts, writing for the court,
observed–in the context of a discussion of the adequacy of that consciousness check–
that “an inmate cannot succeed on an Eighth Amendment claim simply by showing
one more step the State could take as a failsafe for other, independently adequate
measures.” Baze, 553 U.S. at 60-61.
The consciousness checks performed in the four recent Oklahoma executions
are described in part II (F), above. Bearing in mind, as will be discussed in part III,
below, that courts, in method-of-execution challenges, do not sit as boards of inquiry
charged with determining best practices for executions, the court finds that, in the
four recent Oklahoma executions, the consciousness checks, consisting of various
combinations of (i) sternum rubs, (ii) raising eyelids, (iii) speaking loudly, (iv)

36

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 39 of 45

pinching and (v) shaking, were adequate. To be sure, Dr. Van Norman thought a
sternum rub should last “a full 30 seconds,” but Dr. Antognini said that if he “saw
an anesthesia resident, for example, doing a sternal rub for 30 seconds, I would be
very upset with them.” Tr. 602 (Dr. Van Norman), 684 (Dr. Antognini). The court
is not required in this case to sort that out. The evidence gives the court no cause for
concern that the IV Team will fail to perform adequate and effective consciousness
checks.32
I. The plaintiffs’ proposed alternative methods of execution
The foregoing findings relevant to the first Glossip prong are dispositive of
plaintiffs’ claims. However, in the interest of resolving as many issues as possible,
the court will address plaintiffs’ proposed alternative methods of execution (Glossip
second prong).
Plaintiffs proffer the following alternative methods: (i) execution using
fentanyl in combination with pentobarbital or sodium thiopental (compounded or
commercially procured) and (ii) execution by firing squad.
1. Fentanyl plus compounded or commercially procured pentobarbital or
sodium thiopental.
A state cannot be faulted for failing to use lethal injection drugs that it is
unable to procure through good-faith efforts. Bucklew v. Precythe, ___ U.S. ___,
139 S.Ct. 1112, 1125 (2019). As stated in the conclusions of law (part III, below),
the state is not required to use more than ordinary transactional effort to procure
drugs for lethal injection.

32

Notwithstanding plaintiffs’ speculative argument based on the protocol’s provision for revisions
(doc. no. 643, at 75-76), the court is well-satisfied, from the testimony of Department of
Corrections Director Scott Crow and Director of Operations Justin Farris, that there is essentially
no possibility that the department will water down the provision for consciousness checks.

37

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 40 of 45

The credible and uncontradicted testimony of Director of Operations Justin
Farris satisfies the court that the defendants have been unable to procure fentanyl,
pentobarbital or sodium thiopental–for execution purposes–from commercial
sources with ordinary transactional effort. Tr. 1000-04. This comes as no surprise.
See, Glossip, 576 U.S. at 869-70 (discussion of anti-death penalty advocates’
pressure on pharmaceutical companies to refuse to supply lethal execution drugs).
As for compounding, the testimony of Mr. Farris and Dr. Buffington
establishes with equal clarity that the sources suggested by the plaintiffs have been
unwilling, have lacked the requisite technical proficiency, or have been legally
unable to produce compounded drugs for lethal injection purposes. (Mr. Farris: Tr.
1002-03 (dealings with pharmaceutical suppliers), 1004 (contacts with other states),
1004-07 (unsuccessful contacts with compounding labs, including the University of
Oklahoma); Dr. Buffington: 927-929 (compounding and related compliance
requirements), 942-43 (facilities required for compounding), 944-45 (facilities and
materials required for compounding), 948 (search for contract manufacturer of
pentobarbital).) Plaintiffs’ evidence on the issue of compounding as a source of
plaintiffs’ preferred drugs does not go much farther than to suggest, essentially
hypothetically, that compounding could possibly be a way to get the drugs.
Plaintiffs’ evidence stops well short of establishing, in the face of practical and
regulatory hurdles, that compounding is a viable option for the Oklahoma
Department of Corrections.
2. Firing squad.
In evaluating plaintiffs’ proposed firing squad alternative, it must be borne in
mind that analysis under Glossip’s second prong “is a necessarily comparative
exercise.” Bucklew, 139 S.Ct. at 1126 (emphasis in original).

38

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 41 of 45

Plaintiffs called Dr. James Williams to testify in support of the firing squad
alternative. His qualifications to address that subject are impeccable. He clearly
believes that “execution by firing squad is feasible and practicable and it’s
efficacious.” Tr. 326. But Dr. Williams did not, in terms, compare the pain incident
to a firing squad execution with the pain resulting from execution with midazolam
and the other two Chart D drugs. And he did not really go so far as to support
plaintiffs’ contention that execution by firing squad “causes a quick and painless
death.” Doc. no. 594, at 17. Citing Dr. Williams’ testimony, plaintiffs assert that
the “impact of firing squad bullets on a prisoner’s cardiovascular bundle will not
cause the prisoner pain and suffering in the short time before the prisoner loses
consciousness.” Doc. no 643, at 39 (emphasis in original). Accepting that as true
for the sake of discussion, this proposition ignores the fact that the bullets first have
to shatter the inmate’s sternum to get to the cardiovascular bundle, then they exit by
penetrating the inmate’s ribs or spinal column (or both). Tr. 1163 (Dr. Yen). Dr.
Williams testified unequivocally that “sternal fractures in a non-gunshot wound
situation are very painful.” Tr. 346. And he acknowledged that it is “unclear how
much a single gunshot wound to the sternum, let alone multiple gunshot wounds in
the case of firing squad, how much pain that individual would experience in the short
period of time he or she would remain conscious before the heart ceased organized
beating.” Id. After giving that noteworthy disclaimer, Dr. Williams opined that a
person shot in the chest (which the court takes, in context, to have been a reference
to the sternum33) “would not experience anything that most of us would describe as
pain.” Id.

33

If Dr. Williams was not referring here to the sternum, then his only testimony about the pain
incident to a sternal fracture is that it is “very painful.”

39

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 42 of 45

The court is unpersuaded. Though Dr. Williams persuaded the court that
execution by firing squad, done right, is highly likely to cause a quick death, it is
difficult to conceive that a method of execution which requires the sternum to be
shattered as the bullets make their way to the cardiovascular bundle (and on to the
ribs or spinal column) can be called painless. In fairness to Dr. Williams, it should
be noted that, as an expert testifying on behalf of the party with the burden of proof,
he was in the unenviable position of advancing a proposition which, though possibly
meritorious, is inherently difficult to prove. These difficulties with the proposition
plaintiffs seek to advance through Dr. Williams, in combination with the absence of
any explicit attempt to compare the pain of a firing squad execution with the pain
resulting from execution under Chart D, cause the court to reject plaintiffs’ firing
squad alternative.34
III. Conclusions of Law
The Eighth Amendment standards governing this case were articulated by the
Supreme Court in its lethal injection trilogy, Baze v. Rees, 552 U.S. 35 (2008);
Glossip v. Gross, 576 U.S. 863 (2015) (this case); and Bucklew v. Precythe, ___
U.S. ___, 139 S.Ct. 1112 (2019).
“[T]he Eighth Amendment does not guarantee a prisoner a painless death–
something that, of course, isn’t guaranteed to many people, including most victims
of capital crimes.” Bucklew, 139 S.Ct. at 1124. Also worthy of note is the fact that
the Supreme Court “has yet to hold that a State’s method of execution qualifies as
cruel and unusual,” id., and that the deference that is due a state’s choice of execution
procedures means that courts, in method-of-execution challenges, do not sit as

34

Oklahoma statutes provide for execution by firing squad in some circumstances. See, 22 O.S.
§1014. Nothing in this discussion should be understood to shed light one way or the other on the
constitutionality of execution by firing squad.

40

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 43 of 45

“boards of inquiry charged with determining ‘best practices’ for executions.”
Bucklew, at 1125, quoting from Baze, 553 U.S. at 51-52. On this point, it is
noteworthy that, in Baze, the Court declined to attribute constitutional significance
to the fact that the consciousness check was performed “by the warden and deputy
warden through visual inspection,” with no particular level of expertise noted to
qualify those officials to perform that task. Baze, 553 U.S. at 45, 60.
Glossip First Prong. In Bucklew, the court, summarizing its decisions in Baze
and Glossip, held that the inmate who challenges the state’s method of execution
must show that the state’s method presents “a substantial risk of severe pain.”
Bucklew, 139 S.Ct. at 1125.35 A method of execution that presents a “substantial
risk” is one that “is sure or very likely to cause serious illness and needless
suffering.” Glossip, 576 U.S. at 877 (quoting from Baze, emphasis in original).
The inmate must show that “the risk [of infliction of severe pain] is substantial
when compared to the known and available alternatives.” Glossip, 576 U.S. at 878
(internal quotation omitted). This is where the second prong comes in.
Glossip Second Prong. It is clear from Bucklew (again summarizing Baze
and Glossip) that the alternative method of execution the inmate is obliged to
propose must be “feasible and readily implemented,” and it must be one that “the
State has refused to adopt without a legitimate penological reason.” Bucklew, 139
S.Ct. at 1125. To be considered at all, the inmate’s proposal must be “sufficiently
detailed to permit a finding that the State could carry it out relatively easily and
reasonably quickly.” Bucklew, 139 S.Ct. at 1129 (internal quotation omitted). And

35

In In re Ohio Execution Protocol Litig., 946 F.3d 287, 290 (6th Cir. 2019), cert. denied sub nom.
Henness v. DeWine, 141 S.Ct. 7 (2020), the Sixth Circuit appears to have adopted (much to the
consternation of Justice Sotomayor, 141 S.Ct. at 9) the pain associated with hanging as a
benchmark. It is not necessary for this court to determine the merits of that approach.

41

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 44 of 45

the “mere fact that a method of execution might result in some unintended side
effects does not amount to an Eighth Amendment violation.” Glossip, 576 U.S at
882, n. 3. States, of necessity, are free to use previously untried methods. It is quite
unlikely that an untried method will pass muster as the inmate’s proposed
alternative: “[C]hoosing not to be the first to experiment with a new method of
execution is a legitimate reason to reject it.” Bucklew, 139 S.Ct. at 1130.
The proposed alternative need not be one “presently authorized by” state law.
Bucklew¸ 139 S.Ct. at 1128. Thus, the inmate “may point to a well-established
protocol in another State as a potentially viable option.” Id. But it is not enough to
argue for “a slightly or marginally safer alternative.” Glossip, 576 U.S. at 877
(quoting from Baze). The “difference [in risk] must be clear and considerable.”
Bucklew, 139 S.Ct. at 1130. That said, in a passage that has a natural tendency to
accentuate the importance of the first prong (degree of risk and severity of pain), the
Court, in Bucklew, observed that “we see little likelihood that an inmate facing a
serious risk of pain will be unable to identify an available alternative—assuming, of
course, that the inmate is more interested in avoiding unnecessary pain than in
delaying his execution.” Id. at 1128-29.
A state is not required to use more than “ordinary transactional effort” to
obtain drugs for lethal injection. In re Ohio Execution Protocol, 860 F.3d 881, 891
(6th Cir. 2017). “Proof that lethal injection drugs are available with ordinary
transactional effort requires more than mere speculation, more than just a showing
of hypothetical availability.” Middlebrooks v. Parker, 15 F.4th 784, 786 (6th Cir.
2021).
IV. Conclusion
The prerequisites of a successful lethal injection challenge under the Eighth
Amendment have been made clear by the Supreme Court in Baze, Glossip and
Bucklew. The plaintiff inmates have fallen well short of clearing the bar set by the
42

Case 5:14-cv-00665-F Document 648 Filed 06/06/22 Page 45 of 45

Supreme Court. Consequently, the Eighth Amendment, as construed and applied by
the Supreme Court in its lethal injection cases, does not stand in the way of execution
of these Oklahoma inmates under Chart D of the Oklahoma lethal injection protocol.
A final judgment will be entered forthwith.
Dated this 6th day of June, 2022.

14-0665p158 FoF CoL_.docx

43