Advisory Board
Richard Branson MS, RRT, FAARC
Associate Professor of Surgery
University of Cincinnati College of Medicine
Cincinnati, OH
John Eichhorn MD
Professor of Anesthesiology
University of Kentucky
Lexington, KY
Ivan Frantz MD
Professor of Pediatrics
Tufts University School of Medicine
Boston, MA
Nicolette Mininni RN, CCRN, MEd
Advanced Practice Nurse, Critical Care
University of Pittsburgh Medical Center
Pittsburgh, PA
Frank Overdyk MSEE, MD
Professor of Anesthesiology
Medical University of South Carolina
Charleston, SC
M. Terese Verklan PhD, RNC
Associate Professor of Nursing
University of Texas Health Science Center
Houston, TX
Continuing
Education for
Nurses (CE) and
Respiratory
Therapists
(CRCE)
Unrecognized respiratory depression
on the general care floor (GCF),
culminating in respiratory arrest or
“code blue” is a nightly occurrence
at hospitals across the United States. A significant
portion of these respiratory arrests occur in
postoperative patients receiving opioid analgesics
and sedatives, which contribute to respiratory
depression. Failure to recognize respiratory
depression and institute timely resuscitation has
led to cardiopulmonary arrest (often refered to
as “cardiac arrest”), resulting in anoxic brain injuries
and deaths. Herein, we will review the extent
of respiratory depression, its pathophysiology,
and discuss solutions to the problem, which
comes at huge human and economic cost to our
medical system.
Background
Respiratory depression is recognized as a
serious complication of opioid analgesic therapy.
In 2001, the Joint Commission on the Accreditation
of Healthcare Organizations (JCAHO) recommended
more aggressive pain management
in response to scientific data suggesting widespread
undertreatment of pain.1 Not unexpectedly,
the incidence of opioid related adverse drug
events more than doubled in one study examining
the impact of the JCAHO pain therapy standard.
2
Although data concerning the incidence
of opioid-induced respiratory arrest is limited,
patients suffering this condition may indeed
constitute a prominent subgroup in the 350,000
to 750,000 patients suffering an in-hospital cardiac
arrest annually.3 Outcomes in these events
are often catastrophic. Only one in five patients
suffering an in-hospital cardiac arrest survives
to hospital discharge. Patients in “unmonitored”
beds, which include the majority of postsurgical
patients on opioid analgesics, are twice as likely
to receive delayed defibrillation therapy and by
corollary, ventilatory support, than patients in
monitored beds. Patients arresting at night, as
Postoperative
Respiratory Depression
and Opioids
Frank J. Overdyk, MSEE, MD
seen in most case reports of opioid induced respiratory
arrest, are more likely to be “unmonitored”
and have even worse outcomes than patients
arresting during the day, with only a 15%
change of survival until discharge and a 89%
chance of an unfavorable neurologic outcome.4
Fragmented evidence of these devastating adverse
events prompted the Anesthesia Patient
Safety Foundation (APSF) to sponsor a Workshop
on the Dangers of Postoperative Opioids in
the fall of 2006.5
It is important to recognize that there are
a number of factors which predispose postoperative
patients on the GCF to respiratory depression.
Opioid analgesics and sedatives depress
respiratory rate. Residual anesthetics blunt airway
reflexes, lower airway tone, reduce lung capacities,
and disrupt the sleep cycle. Supplemental
oxygen may reduce respiratory drive. Yet the
majority of patients establish a respiratory equilibrium
between factors which depress respiration
(opioids, sedatives, pain) and those which
stimulate it (pain, hypoxic/hypercarbic drive,
pH sensors, airway reflexes) without experienc-
Failure to recognize
respiratory depression and
institute timely resuscitation
has led to cardiopulmonary
arrest resulting in anoxic
brain injuries and deaths.
Medical errors and their impact on patient
morbidity, mortality, and the estimated cost
of care are receiving increased scrutiny—
an area of particular concern for errors
caused by increased demands on physicians and nurses,
particularly from caring for more, and sicker patients.
Although many errors are individualistic, their severity
and associated morbidity may be significantly reduced
through prompt detection. Compounding this problem,
the US is in the midst of a nursing shortage crisis that is
expected to worsen with the aging population. Improving
patient safety through a combination of monitoring
technology, better education of clinicians and refinement
of monitoring protocols can help to attain the goal of zero
tolerance for serious patient harm. Helping to reach this
goal is the purpose of this new serial continuing education
program. In each issue of Initiatives in Safe Patient
Care, we will address topics in patient surveillance in
evidence-based articles, symposia, commentaries and
case studies and each issue will be accredited for continuing
education credits.
In this inaugural issue, Dr. Overdyk addresses the critical
problem of unrecognized respiratory depression (RD) in
the setting of postoperative opioid therapy as a significant
contributor to in-hospital morbidity and mortality. We
have convened a multidisciplinary panel of clinicians to
give a ‘real world’ perspective on postoperative respiratory
depression (RD) and opioid therapy.
Continued on page 4
Enhancing patient safety through improved surveillance
in Safe Patient Care
2
A multidisciplinary panel of clinicians experienced
in the respiratory and pharmacologic
nuances of postoperative, opioid-inclusive pain
management was asked to give a ‘real world’
perspective on postoperative respiratory depression
and the issues presented in this review.
In this discussion, they state their views on the
pervasiveness of the problem, the challenges of
accurate pain and sedation assessment, alternatives
to opioids as postoperative analgesics,
and the role of monitoring technology and clinician
education and awareness in reducing the
incidence of respiratory adverse events.
Is it realistic to have a zero tolerance policy for
opioid-related respiratory depression given the
medical complexity of current in-patients?
Moon: The main problem for in-patient analgesia
is that the most effective systemic therapies
have a very narrow therapeutic margin, and a
risk of respiratory depression. When intermittent
intramuscular dosing regiments were replaced
by patient controlled analgesia (PCA)
over 20 years ago, the hope was that opioid-induced
respiratory depression could be avoided
entirely. That turned out not to be the case. Furthermore,
when patients are weaned from PCA
onto oral narcotics, drug absorption and plasma
levels are much more variable and unpredictable,
and respiratory events continue to occur.
Certain comorbidities, particularly obstructive
sleep apnea, increase the perioperative risk and
specific genotypes also appear to increase the
probability of respiratory depression.1 A zero
tolerance policy for respiratory depression is not
yet achievable, although it should be our ultimate
goal.
Hansen: I believe a zero tolerance policy is possible
with education of the care team, and with
increased monitoring that uses centralized clinical
trending exported to the nurse when clinical
parameters alarm. This is one our top priorities
– for many years we have had standing orders
for preoperative assessment of patients at risk
for sleep apnea or respiratory depression. The
health risk status of our patients is increasing, so
we have heightened our attention to strategies to
reduce the risk. We have also consistently allocated
more capital towards monitoring technology
and treatment devices over the years. Today
we have made a commitment to allocate several
million dollars to expanding our monitoring
network.
Will a single monitoring technology be able to
help us attain a zero tolerance policy, together
with proper communication tools?
Moon: The requirements of any monitoring device
are simplicity, high alarm sensitivity and
specificity, ease of use, and reliability. With these
constraints, respiratory monitoring has traditionally
been based upon measurements of
breathing rate. Despite its apparent simplicity,
accurate monitoring of respiratory rate at the
bedside has been technically difficult in practice.
Expired carbon dioxide, inductance plethysmography
and electrical impedance methods
all have some drawbacks. New technology uses
tracheal sound monitoring to detect breathing,
which is an improvement over techniques
but does not assess gas movement. Whether the
benefits of greater ease of use and reliability will
offset its lack of oxygen or carbon dioxide sensing
remains to be seen. Such technology must
be coupled with algorithms more sophisticated
than a simple rate calculator. The wide variation
in respiratory rate in normal individuals implies
that specific threshold respiratory rate alarms
will generate false positives and negatives. Closer
to ideal would be a respiratory minute volume
monitor, which does not yet exist as a bedside
monitor. Hypoxemia can be a late sign of hypoventilation,
thus pulse oximetry cannot be relied
upon as a single technique New techniques
must therefore be developed, with parameters
such as hemoglobin-oxygen saturation, end-tidal
PCO2, respiratory pattern or sleep state, most
likely in combination.
Would opioid sparing techniques make this problem
sufficiently rare that it would not be an issue
anymore? Are opioids avoidable?
Moon: Opioid sparing techniques today include
use of regional analgesia and nonopioid medications
such as COX inhibitors, acetaminophen,
and anticonvulsants such as pregabalin. Along
with opioids, the use of such techniques probably
reduces the likelihood of respiratory depression,
although this is as yet unproven. Moreover,
in the United States, many systemic adjunctive
agents are only available for enteral use. Ketorolac
is the sole exception, but it has side effects
which make it less than ideal in many postsurgical
patients. There are no parenteral COX-2
inhibitors available in the US, and it is unlikely
that there ever will be. Except for dense regional
block (which is difficult to implement on a general
ward), currently available adjuncts cannot
entirely replace opioids for most surgeries. An
unacceptable incidence of respiratory depression
(i.e. greater than zero) is therefore probably
unavoidable with today’s drugs and monitoring
techniques. With regard to new non-opioid
drug development, the greatest promise perhaps
lies in pharmacotherapy directed at sites other
than the μ-receptor, such as the sodium channel
Nav1.7.2,3
Are current time interval standards adequate to
allow clinicians to assess postoperative patients
on the general care floor?
Jarzyna/Pasero: Monitoring frequency for patients
receiving opioids in the early postoperative
period needs to be more frequent than “every
4 hours” after early vital sign monitoring
criteria are met. The first postoperative 24 hours
Panel Discussion:
Challenges in the Prevention of Postoperative
Respiratory Depression and Opioid Therapy
Moderator: Frank Overdyk, MD
Panelists: Richard Moon, MD;
Harold Oglesby, RRT;
Donna Jarzyna, RN, MSN;
Chris Pasero, MS, RN, FAAN;
Kathryn Hansen, MPH.
The requirements of any
monitoring device are
simplicity, high alarm
sensitivity and specificity,
ease of use, and reliability.
- Moon -
www.initiatives-patientsafety.org 3
seems to be a high-risk period during which
more frequent monitoring is indicated. Changes
in therapy such as dose escalation to meet pain
needs or decreased pain levels without concurrent
adjustment of opioid dosing will prolong
this risk, making a 72 hour period of systematic
monitoring prudent. The establishment of realistic
pain goals together with monitoring of sedation
and respiratory quality will help prevent
situations leading to oversedation and respiratory
depression. Despite similarities among high
risk patients, risk may be practice specific. Identifying
populations at risk is challenging. Facility
specific monitoring of naloxone rescue dosing
and Rapid Response Team data can be incorporated
with the consideration of other known
high risk characteristics such as obesity or a history
of sleep apnea, pulmonary disease, multiple
co-morbidities and the elderly patient. The
implementation of risk-reducing interventions
and the use of opioid sparing multimodal therapy
will ultimately decrease the overall risk for
respiratory depression within the institution.
Oglesby: Our monitoring of PCA patients was
inadequate prior to our move to continuous respiratory
monitoring via EtCO2. Routine monitoring
consisted of vital signs every 30 minutes
x 2, every 1 hour x 4, and then every 4 hours.
There is ample time between these time frames
for patients to deteriorate without recognition.
When a patient demonstrated signs of deterioration,
the first response would often be to place
the patient on O2 that would tend to mask any
underlying hypoventilation. This may have occurred
at the time of the fourth “every 1 hour”
vital sign check, so when the nurse entered the
room to perform the first or second 4-hour interval
check, she or he could very well be unable
to arouse the patient who by then would be
apneic and without a pulse. In an effort to improve
patient safety, our facility added respiratory
therapists (RTs) to help assess patients’ respiratory
status, EtCO2 trends, and SpO2. This
additional monitoring by both nursing and RTs
combined with continuous respiratory monitoring
adds another layer of patient safety.
Are reduced monitoring intervals attainable?
What are the impediments?
Jarzyna/Pasero: Yes, if systems can be tuned
to decrease the nursing time spent in documentation
and if monitoring is focused on the
high-risk patient. Identification of criteria constituting
high risk patients presents numerous
challenges and is an ongoing process. Sedation
and monitoring can be supplemented with pulse
oximetry monitoring in patients not receiving
supplemental oxygen and with end tidal carbon
dioxide monitoring in those who are receiving
supplemental oxygen. This high tech monitoring
should occur for patients at high risk for respiratory
depression and for patients whose sedation
status deteriorates. The expense of high
tech monitoring is an impediment towards wider
use as is the training for general care nurses to
help them distinguish between oxygenation and
ventilation. This must be weighed against the
“expense” of unanticipated adverse events such
as, cardiac or respiratory arrest.
How can technology help nurses and respiratory
therapists to better monitor patients and communicate
with clinical staff? Which particular
assessments and communications are in most
urgent need for improvement?
Jarzyna/Pasero: Nurses must understand how
to integrate technological monitoring data with
pain intervention to improve care. Among the
factors in urgent need of improvement are identifying
and communicating respiratory status
trends that indicate poor ventilation and poorly
controlled pain. Additionally, physician orders
to nurses must allow flexibility for the management
of pain in acutely ill patients. Patients are
frequently oversedated and in pain. When this
occurs, the nurse must be able to reduce the opioid
dose, adjust or discontinue other sedating
medications which cause CNS depression, and
implement nonsedating adjunctive pain treatments.
Hansen: Advanced monitoring of oxygenation
and ventilation is critical, and a centralized station
will complement communication of critical
events to the bedside nurse. Ideally, the communication
of the critical event is transmitted
through a beeper or phone directly to the nurse.
Technology with an algorithm to trend clinical
changes provides the care team with clinical
information that can be used to intervene to
prevent an adverse event. Advanced technology
provides alarms to protect the patient. Rather
than ignoring the alarm, an educated care team
focuses on the clinical assessment required when
monitors alarm. We have expanded our monitoring
to include capnography and oximetry at
a minimum for may of our patients being transferred
from the post anesthesia care unit to the
GCF. Critical to this is the respiratory therapist
who rounds on the patients systematically.
Oglesby: I believe that monitoring patient via
a reliable respiratory monitoring device adds
an extra layer of patient safety. EtCO2/SpO2 in
combination provides an excellent manner in
which to monitor these patients. With these devices
you are able to continuously monitor both
oxygenation and ventilation with minimal interference
with the patient’s daily hospital routine.
With the use of these devices the nurse or respiratory
therapist would be able to give the physician
detail data of the degree of respiratory depression
present. The wise RT would realize the
EtCO2 is not going to be equal to the patient’s
paCO2 and would use the trending of the EtCO2
to guide their assessment. A trend of an increasing
EtCO2 and stable and decreasing SpO2 would
lead to the skilled RT or nurse to believe that the
patient’s respiratory status is worsening and to
suggest an action plan. This trending method
would be the most important assessment being
performed by the nursing and RT, because the
EtCO2 rise will provide and early indication of
respiratory depression in this patient population
versus other monitoring methods.
To what extent does current monitoring interfere
with patient rest, particularly at night? What are
the difficulties of assessing pain, sedation levels,
and respiratory/cardiovascular vital signs at
night?
Jarzyna/Pasero: It is as important for patients
to obtain adequate rest and sleep as it is for
nurses to insure patient safety. The addition of
a noninvasive method of monitoring ventilation
is very tempting to avoid awakening the patient
at night, however, just as counting the rise and
fall of the chest wall does not necessarily indicate
adequate respiration, mechanical monitoring
has flaws and can lead to a false sense of security.
For example, pulse oximetry can suggest
adequate oxygen saturation in patients who are
actively experiencing respiratory depression.
Monitor alarms often awaken patients enough
to remind them to take a breath but do not correct
the problem of respiratory depression. The
use of EtCO2 monitoring, which determines respiratory
rate by registering the exhalation of an
adequate volume of carbon dioxide, may be an
important option for some patients. Again, the
decision to use mechanical monitoring must be
individualized according to risk factors. In addition,
whether or not mechanical monitoring
is used, it is imperative that nurses perform a
thorough respiratory assessment that includes
determining the depth, regularity, noisiness, and
Advanced monitoring of
oxygenation and ventilation
is critical, and a centralized
station will complement
communication of critical
events to the bedside nurse.
- Hansen -
4
rate of respiration, and arouse patients with unacceptable
respiratory status for further evaluation.
The nurse can be empowered to initiate
mechanical monitoring for patients whose respiratory
status appears to be deteriorating.
Oglesby: EtCO2 is a wonderful method of monitoring
patients on PCA therapy, however, there
is the need to deal with alarms and the placing of
a nasal cannula to assess EtCO2 levels. In my experience,
patients who are educated about monitoring
are very willing to wear the device and
tolerate the alarms. Night alarms are associated
with periods of apnea or hypoventilation, so
they are denoting actual events and are not nuisance
alarms. During GCF rounds, it was noted
that patients or the families of patients who tolerated
their devices could state the reason they
were wearing them. Improved alarm systems and
“smart alarms” need to be developed in order to
ensure patient rest while providing a safe environment.
Smart alarms would be able to analyze
a series of input data and interpret that data to
recognize whether the data is significant enough
to warrant an alarm. As for pain assessment—
staff must be educated on how to use the appropriate
pain and sedation scales. It is important to
have a good baseline assessment of the patient’s
risk for sleep apnea prior to the use of any opioid
PCA therapy.
Summary:
Our panel of experts appear to agree that, while
postoperative respiratory depression due to opioids
and sedatives cannot be eliminated, a combination
of sophisticated monitors, including
central surveillance, in addition to better education
of clinicians and refinement of monitoring
protocols, may attain our goal of zero tolerance
for serious patient harm from unrecognized respiratory
depression.
References
1 Romberg RR, Olofsen E, Bijl H, et al. Polymorphism of
mu-opioid receptor gene (OPRM1:c.118A>G) does not
protect against opioid-induced respiratory depression despite
reduced analgesic response. Anesthesiology. 2005;102:522-
30.
2 Cox JJ, Reimann F, Nicholas AK, et al. An SCN9A
channelopathy causes congenital inability to experience
pain. Nature. 2006;444:894-8.
3 Waxman SG. Neurobiology: a channel sets the gain on pain.
Nature. 2006;444:831-2.
ing sequelae. There are two situations which,
when they coexist, transform an effective, uncomplicated,
postoperative analgesic procedure
into one with a catastrophic outcome: (1) the
opioid-treated patient develops decompensated
respiratory depression, in which the equilibrium
described above tilts toward respiratory depression,
and (2) the caregiver fails to recognize respiratory
arrest in time to prevent the irreversible
neurologic and cardiac sequela of prolonged
hypoxia. The challenge for clinicians is to identify
patients whose respiratory compensatory
mechanisms are overwhelmed and respond immediately
to a respiratory arrest should the decompensation
have gone undetected.
Opioid Pharmacology
The potent analgesic effect of opioids is
mediated through the mu (MOR), kappa, and
delta opioid receptors, located throughout the
brain and spinal cord. The respiratory depressive
effect of opioids is also mediated by MOR
receptors, primarily located in the brainstem.
An elegant set of experiments in mice genetically
altered to disable the MOR gene showed how
these two effects are intricately linked.6 Opioids
acted neither as analgesics nor as respiratory
depressants in these animals. Humans display
a polymorphism of the MOR gene located on
chromosome #6 that alters a patient’s sensitivity
to analgesia but not respiratory depression.7
This may explain a recent study in which some
heavily sedated patients in postanesthesia care
complained of significant pain.8
We have shown that pain scores and plasma
opioid levels during patient-controlled analgesia
(PCA) are not linearly related, as one
would logically assume (i.e. higher plasma levels
lead to lower pain scores).9 Thus, there is scientific
evidence to refute a common misconception
in pain management, that is, that a patient
who complains of severe pain cannot be at risk
of severe respiratory depression, presumably because
pain stimulates breathing.
Respiratory depression remains a dangerous
side effect of opioids, regardless of dosing
route, and that includes some of the newer, more
convenient delivery modalities such as transdermal
and intranasal opioid therapy.10 Patients
on chronic opioid therapy who have developed
pharmacologic tolerance present a very challenging
postoperative pain management problem.
Equianalgesic opioid dosing can vary 20- to
100-fold in these patients.
Studies suggest that opioid receptors may
be acutely sensitized by potent intraoperative
opioids such as remifentanil, leading to opioidinduced
hyperalgesia (OIH), a condition similar
to opioid tolerance.11 Thus it is clear that relationship
between the dose of an opioid and its
pharmacologic effects on analgesia and respira-
Postoperative Respiratory Depression
and Opioids — Continued from page 1
Donna Jarzyna, RN-C, MS, CNS is a certified pain
nurse, and works at the University Medical Center
(UMC), University of Arizona, Tucson, as an adult health
clinical nurse specialist for acute pain. She functions
as clinical liaison between the Acute Pain Service and
UMC Nursing Services, and is involved in direct care and
coordination for all patients with neuraxial analgesia,
and consults with patients in intractable pain. She is
also responsible for staff education, quality improvement,
policy and protocol revisions, and research in
pain management. Ms. Jarzyna is actively involved in
the pain management research and has presented many
lectures on the topic at medical meetings. She lives in
Tucson, Arizona.
Richard E. Moon, BSc, MD, CM, MSc, FRCPC, FACP,
FCCP. is Professor of Anesthesiology and Pulmonary
and Critical Care Medicine at the Duke University
Medical Center, Durham, North Carolina. He is the
author or coauthor of dozens of journal articles, book
chapters, and meeting abstracts. His primary research
interests include pulmonary gas exchange under anesthesia,
environmental physiology, pathophysiology
of neurological decompression sickness, mechanisms
of hyperbaric oxygen therapy, mechanisms of postoperative
pulmonary dysfunction, among others. He is
currently Chairman of the Board, Research Foundation,
Undersea and Hyperbaric Medical Society, and Chairman,
DCI Adjunctive Therapy Committee, Undersea and
Hyperbaric Medical Society.
Harold Oglesby, RRT is a registered respiratory therapist
and Manager of the Center for Pulmonary Health at
St. Joseph’s/Candler Health System, Savannah , Georgia.
He is also a member of the Respiratory Advisory Board,
Georgia State Composite Medical Board, Atlanta.
In addition to being a hospital clinical instructor at
Armstrong Atlantic State University in Savannah, Mr.
Oglesby has participated in several publications and
presentations on the use of respiratory care technology.
He lives in Savannah, Georgia.
Frank J. Overdyk, MSEE, MD is Professor of Anesthesiology
and Perioperative Medicine at the Medical
University of South Carolina, Charleston, South Carolina.
Dr. Overdyk is also author or coauthor of dozens
of journal articles, book chapters, and abstracts, and
also writes for the medical press. He is a member of 10
scientific and professional societies and is an editorial
reviewer of 7 medical journals. His research interests
focus on the use of new techologies in anesthesiology.
Dr. Overdyk lives in John’s Island, South Carolina.
Chris Pasero, MS, RN-BC, FAAN is a pain management
author, educator, and clinical consultant from El Dorado
Hills, California. She is a co-founder and past president
of the American Society for Pain Management Nursing
and serves on the Board of Directors of the American
Chronic Pain Association. Ms. Pasero is a Fellow in the
American Academy of Nursing, board certified in Pain
Management Nursing, and the recipient of numerous
pain management clinical practice, journalistic, and
teaching awards. She serves on the Editorial Boards
for Nursing Consult, Federal Practitioner, Pain Management
Nursing, and the Journal of PeriAnesthesia
Nursing. Major publications include numerous pain
management articles, position papers, guidelines, and
book chapters.
Kathryn Hansen, BS, CPC, REEGT is the director of
the Sleep Wellness Center at Saint Joseph Healthcare
in Lexington, KY. She serves on numerous interdisciplinary
committees within the healthcare community
to establish readiness for Joint Commission audits,
enhance education tools on pain management and
facilitates education programs on a proactive method to
treat sleep apnea in postoperative patients. Ms. Hansen
is actively involved in the effects of sleep deprivation
and lectured extensively on this topic.
www.initiatives-patientsafety.org 5
tory depression is complex, and is subject to differences
in genetics, gender, age, comorbidities,
comedications, and dosing route, among other
factors.
Most clinicians prescribing opioids are
unaware of the many nuances in opioid pharmacology,
such as women initially requiring
more morphine than men to attain adequate
pain relief. (This is due to slower equilibration.)
Yet morphine is more potent in women than in
men, which may make them more vulnerable to
severe respiratory depression.12
The prevailing postoperative opioid dosing
scheme in adults starts with a non-weightbased,
“one size fits all” dosing scheme, and a
“titrate to effect” order. However, the safety of
this procedure is heavily dependent on frequent
and responsive patient monitoring on the GCF,
if potentially catastrophic events are to be prevented.
Analysis of Opioid-Induced Respiratory
Arrest
In a high-risk field such as medicine, a catastrophic
event is usually the result of a confluence
of smaller missteps that, in aggregate, lead
to the event, in this case, a respiratory arrest. Every
patient relies on the flawless execution of a
series of steps to ensure that their postoperative
pain therapy is given safely. Patients may tolerate
a mistake in one or two of these steps, but
a breakdown in more than these places them
at great risk of a catastrophic event, especially
with drugs such as opioids, which have a narrow
therapeutic index. The various types of errors
that occur with opioid administration are
well documented in the literature.13 Almost half
of all deaths attributed to medication errors involve
opiates.14 The following list describes some
of the factors in opioid therapy where mistakes
can lead to an adverse event. Vignettes from actual
cases are included.
Coexisting conditions and medications:
Obstructive Sleep Apnea (OSA): Patients with
OSA have an increased risk of postoperative adverse
respiratory events.15 Tools to help identify
patients at risk for OSA are being developed.16
Case event: A patient has a documented history
of OSA, confirmed by polysomnography, but is
not obese, which is a widely recognized risk factor
for OSA. His OSA diagnosis is missed during
preoperative evaluation. Patient suffers respiratory
arrest at 5:00 AM on the day after surgery.
Difficult pain management: Certain patient
populations are likely to have unsatisfactory pain
management with standard dosing regimens.
These include opioid-tolerant patients, including
those who abuse controlled substances.
Case event: Nurses have difficulty controlling
pain in a patient on high-dose, preoperative opioids,
and receive a late night, verbal order for
additional pain medication, outside of clinical
guidelines. Code blue called at 6:00 AM.
Case event: A patient with a known history of
substance abuse receives a fentanyl patch postoperatively.
He suffers a respiratory arrest in the
early morning hours. There is suspicion that he
tampered with the time controlled drug release
mechanisms of the patches.
The elderly: The elderly have been documented
to be more prone to respiratory depression,
largely due to pharmacokinetic and pharmacodynamic
variability of opioids.
Drug interactions: Many physicians are not
aware of drug interactions that increase the risk
of respiratory depression, such as opioids given
with sedatives or certain antibiotics.
System and Communication deficiencies:
Drug ordering: The drug ordering and administration
protocols in many hospitals are still
paper-based and involve multiple data transfers
between multiple individuals. Handwritten orders
are prone to misinterpretation. Many of the
names of the drugs are similar.
Case event: “Discontinue PCA” written at 9:00
AM. Oral (p.o.) opioids ordered. Unit clerk processes
orders at noon, but mistakenly forgets to
remove PCA order from medicine administration
record. Nursing aide administers p.o. opioids
while PCA is still in place. Respiratory arrest
at 6:00 PM.
Case event: Verbal order for morphine is mistakenly
written as “hydromorphone,” a drug 8
times more potent than morphine. Respiratory
arrest follows.
Human and Technical Factors:
Technical: Programming sequence of PCA
pumps has potential for dosing errors due to
confusion with drug nomenclature or dosing
units. Pumps rely on unobstructed flow. Manufacturers
have improved pump user interfaces
and mechanics since many of these problems
came to light.
Case event: Patient on fluid restriction suffers
respiratory arrest after concentration of opioid
is increased 10–fold to minimize fluid administration,
without decreasing infusion rate by 10-
fold.
Case event: Patient’s IV tube is clamped, yet
patient continues to press button and doses accumulate
in tubing. Bolus of opioid is released
upon unclamping.
Human Factors: PCA relies on the patient to
press the pump button and self administer pain
medication. Other family members, nurses, and
strangers have been known to push the PCA
button.
Case event: Mother presses pain button after
troublesome early recovery to ensure “a good
night’s sleep.” Son suffers respiratory arrest two
hours later.
Monitoring:
Current GCF monitoring standards call
for vital signs to be recorded as infrequently as
once per hour on the day of surgery and once
every 4 hours on the second postoperative night.
Manually recorded respiratory rates are notoriously
inaccurate.17 Assessing the level of sedation,
a vital component of opioid monitoring, is
difficult at times, especially at night. Differentiating
a patient who is sleeping comfortably from
a patient who is narcotized with respiratory depression
and at risk of an imminent respiratory
arrest can be challenging.
Case event: A patient’s wife goes to the
nursing desk at 4:00 AM to alert the nurses that
“something is wrong” with her husband. She
tells them he never sleeps on his back and his
snoring is “different.” The nurse checks on the
patient and states he “is resting comfortably
and needs quiet.” He suffers a respiratory arrest
shortly thereafter.
Are Opioids Avoidable?
Given the inherent risk of a serious respiratory
complication with opioids, one approach
is to minimize the use and/or dosage of opioids.
Effective strategies include minimally invasive
surgery with local anesthesia, regional anesthetic
blocks, and non-opioid analgesics, such as nonsteroidal
anti-inflammatory drugs (NSAIDS)
and COX inhibitors. Unfortunately, the effect on
platelet function makes many surgeons wary of
using NSAIDS and COX inhibitors postoperatively.
Drugs that block the N-methyl-D-aspartate
(NMDA) receptor, such as ketamine, dextromethorphan,
or gabapentin look promising
in blocking OHI. However, it is unlikely in the
near term that opioids will be replaced as the
mainstay of in-hospital, postoperative pain therapy,
and therefore, severe respiratory depression
will remain a risk. As of 2006, there were over
10 million inpatient surgeries performed in U.S.
hospitals.18 It can reasonably be expected that a
significant proportion of these procedures will
require parenteral opioids for at least the first 24
hours postoperatively.
Monitoring
Given the unpredictability of a patient’s
response to a given opioid regimen, and the
inherent human factor and design shortcomings
of our drug administration process, careful
monitoring of patients postoperatively on the
GCF is the best approach to prevent catastrophic
adverse respiratory events. In a small but significant
sample of such cases in the American Society
of Anesthesiologists Closed Claims database,
at least half of the events would have been preventable
with better monitoring.19
Extremely effective monitoring solutions
have been developed for the operating room,
such as oximetry and capnography, and they
have greatly reduced the morbidity and mortality
of anesthesia. Unfortunately, monitoring on
the GCF is not as straightforward as in the operating
room (OR), where patients are anesthetized,
immobile, under controlled ventilation,
and continually monitored by a dedicated and
highly trained anesthesia providers. High-risk
6
patients on the GCF are often monitored with
an intermittent or continuous monitoring bedside
pulse oximeter, set to emit an audible alarm
triggered by a threshold value, such as a respiratory
rate less than 8 bpm. Although the monitor
may be identical to that used in the OR, its efficacy
in this setting is much reduced. With five or
more patients to monitor in different locations,
a GCF nurse may note the SpO2 at most a couple
of times per hour and only intervene when the
audible alarm goes off. A written value may be
charted only every 4 hours. Barring an acute inciting
event such as a pulmonary embolus, respiratory
depression that culminates in respiratory
arrest is an insidious, gradual event, which will
escape the notice of the casual observer of intermittent
vital signs.
Spot checks of ventilatory parameters
such as respiratory rate, SpO2, or ETCO2 may
miss the gradual deterioration of ventilatory
efficiency. The reason why the incidence of respiratory
depression has been underestimated
in the literature is because many of the studies
rely on intermittent monitoring.20 Continuous
monitoring, combined with trend analysis and
interpretation, will likely detect a patient about
to cross the threshold from stable respiratory depression
to respiratory decompensation and arrest.
Furthermore, the GCF patient’s wide range
in level of consciousness and mobility present a
new set of challenges that preclude simply installing
operating room monitors on the GCF.
Patients sleep, eat, talk, and go to the bathroom.
Is the drop in oxygen saturation due to hypoventilation
or an artifact from repetitive motion? Is
the drop in end tidal carbon dioxide due to airway
obstruction or due to the patient coughing?
Is the patient asleep or are they so heavily
narcotized from their elevated PaCO2 that they
may arrest at any moment? Why does the respiration
rate alarm sound every 5 minutes and
then silence itself, or was there even someone
there to hear it? These questions suggest that
simply monitoring a single physiologic parameter
and setting an alarm when it reaches a particular
threshold may not be sufficiently specific
or sensitive to detect decompensated respiratory
depression on the GCF. In other words, there
may be too many false positives or false negatives
with single parameter monitoring.
It is more likely a combination of technologies
will be needed for the GCF. Existing, noninvasive
monitoring technologies such as capnography,
oximetry, and transcutaneous CO2,
are being refined and improved to address the
challenges of the GCF. The capnograph detects
and quantifies exhaled CO2 (ETCO2), and can
measure respiration rate. Although ETCO2 can
vary greatly and is subject to artifacts in nonintubated
patients, we have shown that trend
analysis of the respiration rate and ETCO2 using
heuristic algorithms can identify the insidious
onset of opioid-induced respiratory depression
over many hours.21 (Figure 1). Smart algorithms
are being developed which reduce the number
of false positive alarms. In selected GCF patients,
we were able to identify transient airway
obstruction, manifested by snoring, and correlate
the capnograph with different levels of consciousness.
22 (Figure 2). Moon and colleagues
have used the breath-to-breath intervals of the
capnograph to quantify ventilation stability.23
There is evidence that high resolution pulse oximetry
can detect patterns consistent with transient
airway obstruction in patients with OSA,
and may provide an early warning of patients at
risk for respiratory decompensation.24 Transcutaneous
CO2 (PtcCO2) monitors have also been
successfully applied to postoperative patients receiving
opioids.25 Bioacoustic technology, which
uses a microphone to record airway sounds consistent
with breathing, has been used to measure
respiration rate during conscious sedation, and
may have application on the GCF.26 A neurologic
monitor that can help assess level of consciousness
may be a useful adjunct for nurses at times
when the patient needs to be assessed without
being disturbed. Processed electroencephalographs,
used in the operating room to monitor
depth of anesthesia, have been evaluated in their
ability to differentiate stages of natural sleep,
with equivocal results.27
However, each of these technologies have
shortcomings at this time that make them less
than ideal for the GCF patient. The most encumbering
of these is that these instruments still
require tethering of the transducer attached to
the patient to the bedside monitor by means of
a cable or cannula. This limits patient mobility,
satisfaction, and compliance, especially if multiple
monitors are used. Miniaturization of devic-
Figure 2 Continuous, condensed capnograph from a patient demonstrating differing CO2 patterns when snoring
(partial airway obstruction), when sleeping without snoring, and when awake, and while awakening.
Figure 1. Continuous capnography tracing of patient receiving morphine via patient controlled analgesia (PCA). (A)
Gradual decrease in RR after bolus doses of morphine. Pharmacokinetic model shows concentration of morphine at
effect site (brain) increasing. ETCO2 increases as RR falls (B). After opioid level peaks, RR increases gradually (C) while
ETCO2 levels fall (D).
www.initiatives-patientsafety.org 7
es will likely improve this problem.28 False positive
alarms, more common when monitoring
an awake, moving patient, are also a significant
problem with current monitors.29 Unlike the
situation in the OR, where a dedicated provider
immediately assesses the alarm and silences it
if deemed to be a false positive, bedside audible
alarms are a nuisance to the patient and may fail
to alert a provider who is not in the room. Patients,
family members, and at times nurses, disable
the alarm for this reason.
Application of wireless technology to enable
instant, remote notification of care providers,
or centralized respiratory monitoring on the
GCF are excellent solutions. As in centralized
cardiac telemetry, centralized respiratory monitoring
would provide surveillance by an individual
specifically trained in detecting respiratory
depression and interpreting the more sophisticated
algorithms in future respiratory monitors.
It would also relieve nurses from having to silence
bedside alarms that are clearly an artifact.
Commercially available central monitoring applications
for oximetry which use the paging
system for notification of adverse events are OxiNet
®, (figure 1)and Patient SafetyNet. (masimo.
com) One pain pump manufacturer has logically
incorporated oxygenation and ventilation
monitoring into their pump platform, allowing
for an immediate halt to the infusion should
monitoring data suggest respiratory insufficiency.
(cardinalhealth.com) Lastly, the adoption of
a medical device “plug and play” (MDPnP) standard
by medical equipment manufacturers will
allow monitoring data to be disseminated more
readily, and help reduce communication mishaps
with critical data, which are common during
respiratory adverse events. (mdpnp.org)