Applied Nursing Research 55 (2020) 151284

Contents lists available at ScienceDirect

Applied Nursing Research

journal homepage: www.elsevier.com/locate/apnr

Iatrogenic withdrawal syndrome in specialty pediatric critical care T

a,⁎ b
Paula Conrad (MSN, RN, CNL, CCRN, CPN) , Shannon Meyer (MPH) ,
Jon Whiting (MSN, RN, CCRN)b, Jean Anne Connor (PhD, RN, CPNP, FAAN)c,d
a
Department of Infection Prevention and Control, 300 Longwood Ave., Boston Children's Hospital, Boston, MA 02115, USA
b
Cardiovascular and Critical Care Patient Services, 300 Longwood Ave., Boston Children's Hospital, Boston, MA 02115, USA
c
Cardiovascular and Critical Care Patient Services, Galleria 257, 350 Longwood Ave., Boston Children's Hospital, Boston, MA 02115, USA
d
Harvard Medical School, Boston, MA, USA

A B S T R A C T

Aim: To describe the occurrence of opioid and benzodiazepine withdrawal

symptoms in a cohort of pediatric intensive care unit (PICU) patients, the characteristics of this group, and patterns of withdrawal scoring observed during medi-
cation weaning.
Background: Patients in the PICU are a complex and vulnerable population. Opioids and benzodiazepines are routinely administered in this setting. Providers must be
equipped to recognize and assess symptoms of narcotic and benzodiazepine withdrawal.
Methods: A retrospective chart review was conducted to describe all patients admitted to the medical intensive care unit who received continuous infusions of
morphine and midazolam during a one-year period. Patient demographics, diagnosis, and presence of co-morbidities were abstracted. The number of days on
continuous infusions was measured, along with Withdrawal Assessment Tool-1 (WAT-1) scores and documented symptoms that could be associated with withdrawal.
WAT-1 scoring ranges from 0 to 12, a WAT-1 score of 3 or higher is considered to indicate clinically significant withdrawal symptoms. Descriptive statistics were
utilized to summarize demographic and clinical variables.
Results: Among 60 cases, patient ages ranged 5 weeks to 29 years (median 3.5 years). Eighty percent of patients had a primary respiratory diagnosis and 88.3% had
one or more co-morbidities. Forty-four patients (73.3%) had symptoms consistent with withdrawal. Thirty-one percent of patients had a maximum WAT-1 score
between 3 and 8. The majority of patients (55%) had a history of opioid and/or benzodiazepine exposure.
Conclusions: The information learned highlights the need for ongoing conversation and continued study of how best to assess and manage withdrawal syndrome in
pediatric critical care environments.

1. Introduction complex. Medically complex children, defined as having an underlying

medical condition, may have functional limitations and are often de-
Adequate pain management and optimal sedation are fundamental pendent upon medical technology (Shapiro, Henderson, Hutton, & Boss,
aspects of pediatric critical care. Children in the intensive care unit 2017). Medically complex pediatric patients account for a large pro-
receive opioids and benzodiazepines for a multitude of reasons; these portion of healthcare resource utilization, as many experience frequent
include pain control, reduction of anxiety, maintenance of invasive hospitalizations requiring ICU care (Shapiro et al., 2017).
devices such as endotracheal tubes, and optimization of mechanical In the United States, some combination of benzodiazepines and
ventilation (Anand et al., 2010). The use of these agents, while vital to opioids are used in the majority of long-term pediatric ICU (PICU)
providing the necessary care, can result in the development of tolerance patients (Birchley, 2009). The most commonly used drugs to achieve
and withdrawal symptoms. Management of these issues in children analgesia and sedation are morphine, midazolam, fentanyl, and lor-
adds to the existing complexity of pediatric critical care. azepam, but surprisingly there is little agreement regarding the optimal
route and dosing in the pediatric population (Anand et al., 2010). Many
2. Background critically ill patients require opioids and benzodiazepines for an ex-
tended period, which may result in opioid tolerance. Tolerance is de-
The pediatric critical care patient population is rapidly increasing in fined as a decrease in a drug's ability to produce a given effect, or a need
both acuity and complexity. More than 50% of pediatric patients ad- to increase the dosage of a drug to produce the same effect (Ista, Van
mitted to the intensive care unit (ICU) are described as medically Dijk, Gamel, Tibboel, & De Hoog, 2007, p. 1396). Possible risk factors

⁎
Corresponding author at: 300 Longwood Ave, Boston, MA 02115, USA.
E-mail addresses: [email protected] (P. Conrad), [email protected] (S. Meyer),
[email protected] (J. Whiting), [email protected] (J.A. Connor).

https://doi.org/10.1016/j.apnr.2020.151284
Received 17 January 2020; Received in revised form 1 May 2020; Accepted 7 May 2020
0897-1897/ © 2020 Elsevier Inc. All rights reserved.
P. Conrad, et al. Applied Nursing Research 55 (2020) 151284

for the development of tolerance include duration of therapy greater Sedation Withdrawal Scale (SWS), the Sophia Observation withdrawal
than 72 h, gender, and use of short-acting opioids (Anand et al., 2010). Symptoms-scale (SOS), and the Withdrawal Assessment Tool-1 (WAT-1)
Withdrawal syndromes are described as signs and symptoms that (Madden et al., 2017). The SOS tool includes 15 items to measure
develop with the sudden discontinuation of a sedative or analgesic in a opioid and benzodiazepine withdrawal, including heart rate and re-
patient with physical tolerance; the exact symptoms described are spiratory rate, fever, sweating, tremor, motor disturbances, crying, and
variable and may be affected by age, medical condition, and other gastrointestinal symptoms. Psychometric testing demonstrated a sensi-
factors (Tobias, 2000). Initially, much of what was known about tivity of 83% and a specificity of 95% (Ista, De Hoog, Tibboel,
withdrawal syndromes was found in adult literature, or in the context of Duivenvoorden, & Van Dijk, 2013). The WAT-1 tool utilizes 11 items to
neonates born to opioid-addicted mothers. Beginning in the 1990s, ia- quantify IWS, including fever, gastrointestinal symptoms, sweating,
trogenic withdrawal syndromes (IWS) were increasingly described in response to stimulus, muscle tone, and movement disturbances. Initial
PICU patients exposed to opioids and benzodiazepines (Birchley, 2009). testing of the tool demonstrated 87% sensitivity and 88% specificity
This recognition of IWS in the PICU population has resulted in in- with tool validity and generalizability further demonstrated in a follow
creased discussion, review of the literature, and clinical research in the up multicenter study (Franck, Scoppettuolo, Wypij, & Curley, 2012).
last three decades. Both WAT-1 studies were conducted with patients being ventilated for
pneumonia, and most patients were less than 6 years of age. No in-
2.1. Incidence of IWS in pediatric patients formation is provided regarding previous exposure to opioids or ben-
zodiazepines in the study groups. In addition, the role of concurrent co-
While there is widespread agreement that IWS is an important issue morbidities or chronic critical illness is not explored (Franck et al.,
in pediatric critical care, defining the exact prevalence of withdrawal 2012).
syndromes in PICU patients is a challenge. Birchley (2009) estimates Important progress has been made in tool development for the
that approximately one-fifth of PICU admissions result in withdrawal measurement of IWS in pediatric critical care patients. Pediatric critical
syndromes. In one study, 50% of PICU patients receiving continuous care is a dynamic, rapidly changing environment with increasing pa-
infusions of midazolam and fentanyl for 48 h developed withdrawal tient complexity and acuity. Due to capacity demands and decreased
symptoms; this increased to 80% when infusions continued for greater community resources, many critical care patients have extended stays
than five days (Fernandez-Carrion et al., 2013). In an observational in the PICU. These factors have resulted in a changing population in
multicenter study, withdrawal syndrome was recognized in 64% of many pediatric critical care units. While tools exist to measure IWS in
PICU patients (n = 113) receiving sedation and analgesia for at least pediatric critical care patients, accurate assessment of withdrawal can
five days (Amigoni et al., 2017). A prospective, descriptive study of be a challenge. The purpose of this study was to describe the occurrence
twenty-five pediatric patients receiving opioids for more than five days of IWS and related symptoms in a cohort of PICU patients, the char-
found that 45% of these patients exhibited withdrawal symptoms acteristics of this group of patients, and patterns of WAT-1 scoring
(Fisher, Grap, Younger, Ameringer, & Elswick, 2013). Despite the dif- observed during medication weaning.
ficulty in determining the true prevalence of IWS, pediatric critical care
patients are clearly at high risk. 3. Methods

2.2. Symptoms of opioid and benzodiazepine withdrawal 3.1. Clinical setting

The signs and symptoms of opioid and benzodiazepine withdrawal The setting for this study was a large, urban, quaternary, free-
are well-described in the pediatric critical care literature, although most standing children's hospital in the Northeastern United States. The
sources describe opioid-associated symptoms more comprehensively. hospital achieved Magnet® recognition in 2008, and has received
Despite this, it is thought that the differences in symptoms between the Magnet® re-designation in 2012 and 2017. The medical intensive care
drug classes are minimal (Ista et al., 2007). Observed signs and symp- unit (MICU) is one of four ICUs in the institution. The 22-bed MICU
toms develop as a result of central nervous system overstimulation, serves patients with a wide range of medical conditions from birth
hyperactivity of the sympathetic nervous system, and gastrointestinal through young adulthood. Common diagnoses include pulmonary dis-
disturbance (Tobias, 2000). Over twenty-five symptoms are described ease, metabolic disorders, neuromuscular disease, sepsis, and seizure
in the literature, including, but not limited to: tachycardia, hyperten- disorder. Like other PICUs, patients in the unit routinely receive opioids
sion, sweating, fever, tachypnea, vomiting, diarrhea, irritability, tre- and/or benzodiazepines during their stay. When this therapy is being
mors, yawning, and sneezing (Tobias, 2000). discontinued, the MICU utilizes the WAT-1 tool to measure withdrawal
symptoms. WAT-1 scoring is completed a minimum of once per 12-hour
2.3. Assessment of withdrawal shift and is documented in the electronic health record and commu-
nicated during patient rounds. The WAT-1 has four sections that are
It is vital that intensive care providers assess and recognize the completed: 1. Completed via chart review - any loose/water stools, any
symptoms of narcotic and benzodiazepine withdrawal. This recognition vomiting/wretching/gagging, temperature greater than 37.8 °C; 2. 2-
can be particularly challenging in the pediatric population because the minute pre-stimulus observation – state, tremor, any sweating, un-
symptoms may be confused with signs of inadequate sedation (Ista coordinated/repetitive movement, yawning or sneezing; 3. 1-minute
et al., 2007), as well as conditions such as delirium or anticholinergic stimulus observation – startle to touch, muscle tone; 4. post-stimulus
toxicity (Madden, Burns, & Tasker, 2017). When evaluating a patient recovery – time to gain calm state. An aggregate score from 0 to 11 is
for possible IWS, clinicians must rule out all possible causes of pa- assigned. Nurses will routinely obtain more frequent WAT-1 scores if
thology, including sepsis, neurologic conditions, or cardiovascular dis- there is a concern for IWS or if rescue doses of medication are required
ease (Whelan, Heckmann, Lincoln, & Hamilton, 2015). to treat IWS symptoms. A WAT-1 score of 3 or higher is considered to
The pediatric critical care nurse is often the first clinician to detect indicate clinically significant withdrawal symptoms (Franck et al.,
the presence of withdrawal and therefore must possess the knowledge 2012).
and tools to recognize and describe the signs and symptoms he or she
observes (Ista et al., 2007). Early tools developed to assess withdrawal 3.2. Data collection
focused on the neonatal population. Since then, additional tools have
been designed and tested for use in older children (Ista et al., 2007). Following hospital Institutional Review Board approval, a list was
Commonly used tools to measure IWS in pediatric patients include the obtained from the hospital's pharmacy informatics team including all

2
P. Conrad, et al. Applied Nursing Research 55 (2020) 151284

Table 1 7.3, while the median was 4 days. This differs somewhat from the pa-
Patient characteristics. tients in the most recent WAT-1 validation study, where the medians
n = 60 were 6 and 7 days for benzodiazepines and opioids, respectively
(Franck et al., 2012). Treatment for 5 or more days is widely considered
Age to be a risk factor for IWS (Best, Boullata, & Curley, 2015). Thirty (50%)
Mean 8.3 years
of the patients in the MICU cohort would be considered at high risk for
Median 3.5 years
Most frequent 3 years
withdrawal based upon this accepted risk factor.
Range 5 weeks–29 years In the MICU population studied, 31.7% were found to have a
Gender n (%) maximum WAT-1 score between 3 and 8. A score of 3 or more is con-
Female 34 (56.7%) sistent with opioid withdrawal (Whelan et al., 2015). It is important to
Male 26 (43.3%)
note, however, that 44 (73%) of the MICU patients studied had docu-
Respiratory-related diagnosis 48 (80%)
Presence of comorbidities 53 (88.3%) mented symptoms associated with opioid and benzodiazepine with-
drawal. Of the 44 patients with documented symptoms, 25 (56.8%)
were weaned, 14 (31.8%) were not, and the remaining did not have
MICU patients who received continuous infusions of morphine and documented weaning information. This points to a possible gap be-
midazolam during a one-year time frame (n = 60). A retrospective tween what is observed by nurses at the bedside and what is captured
chart review was conducted to describe patient demographics, diag- by the WAT-1 tool. While the WAT-1 tool has performed well in validity
nosis, and presence of co-morbidities. The number of days on con- testing, there is some concern that it may not measure benzodiazepine
tinuous infusions was measured, as well as WAT-1 scores and docu- withdrawal as well as it measures opioid withdrawal (Whelan et al.,
mented symptoms that could be associated with withdrawal. Other 2015).
factors such as history of opioid and benzodiazepine exposure were also This retrospective chart review also provides a description of patient
noted. Descriptive statistics were utilized to summarize demographic characteristics that are important to consider. While 80% of the MICU
and clinical variables. patients presented with a primary respiratory diagnosis, 88.3% of pa-
tients also exhibited one or more co-morbidities. This illustrates the
complexity of the patient population served in the MICU. The WAT-1
4. Results
tool was tested primarily on patients less than 2 years of age who were
receiving mechanical ventilation for bronchiolitis or pneumonia
Among the 60 cases, patient ages ranged from 5 weeks to 29 years,
(Franck et al., 2012). The investigators did not include information on
with a median age of 3.5 years. Gender distribution was 56.7% female
patient comorbidities in the study population. Patients with multiple
and 43.3% male. Eighty percent of patients had a primary respiratory
complex medical problems often exhibit symptoms that are similar to
diagnosis and 88.3% of patients had one or more associated co-mor-
those seen in IWS at their baseline and may be treated with opioids or
bidities (Table 1). Length of continuous opioid and benzodiazepine
benzodiazepines for associated medical conditions. This could confound
infusions ranged from 1 to 36 days; the mean was 7.3 days and the
assessment of IWS in this population. Many of the MICU patients (55%)
median 4 days. A statistically significant difference was found between
had a previous history of opioid and/or benzodiazepine exposure. In-
patients who showed signs and symptoms of withdrawal, with a mean
formation regarding previous exposure to these medications was not
of 8.6 days on continuous opioid and benzodiazepine infusions, and
provided in the WAT-1 validation studies. It is possible that patients
those who did not show any signs or symptoms of withdrawal, with a
who receive these medications could have a different response and
mean of 2.9 days (p = 0.001) (Table 2). Forty-four patients (73.3%)
experience withdrawal symptoms differently.
had documented symptoms consistent with IWS during medication
weaning or after discontinuation. The most commonly observed
symptoms were gastrointestinal (50%), followed by cardiovascular 5.1. Limitations
symptoms (46.7%) and irritability/agitation (43.3%) (Fig. 1). Fifteen
percent of patients had a maximum WAT-1 score of 1 to 2, while 31.7% As this was a retrospective review, data regarding prior opioid or
had a maximum score between 3 and 8 (Table 3). The majority of pa- benzodiazepine exposure was available for 73% of patients with the
tients (55%) had a history of opioid and/or benzodiazepine exposure, remaining 27% of patients noted to be limited or missing. The review of
while 18.3% had no documented history of previous exposure (Fig. 2). cases was descriptive in nature, limiting the ability to conclude that
Those with previous exposure to opioids and/or benzodiazepines in- there were relationships between patient characteristics and the oc-
cluded patients with previous ICU hospitalizations, as well as patients currence of IWS, while highlighting an opportunity for future study.
requiring these classes of medications for surgical procedures or un- Given this is a review of a single ICU, caution must be employed when
derlying medical conditions, such as seizure disorder. considering generalizability to other pediatric MICUs.

5. Discussion 5.2. Implications for nursing practice and future inquiry

The cohort studied here was representative of the MICU population The recent increase in research and dissemination on the topic of
regarding gender distribution, age range, and types of diagnoses. The pediatric withdrawal syndrome indicates the importance of this issue in
mean number of days that the group received continuous infusions was critical care. As the pediatric intensive care environment continues to

Table 2
Days on continuous infusion of morphine and midazolam.
All participants Showed withdrawal signs or symptoms Did not show withdrawal signs or symptoms p-Value
n = 60 n = 44a n = 12a

Mean number of days 7.3 days 8.6 days 2.9 days p = 0.001
Median number of days 4 days 5.5 days 1.5 days
Range of days 1–46 days 1–46 days 1–9 days

a
Data missing for n = 4 participants on “withdrawal signs or symptoms”.

3
P. Conrad, et al. Applied Nursing Research 55 (2020) 151284

Withdrawal Signs/Symptoms
80

70

60

Percent (%) 50

40

30

20

10

Fig. 1. Withdrawal signs/symptoms.

Table 3 6. Conclusion
WAT-1 scoring.
Max WAT-1 score N (%) Clinicians in the pediatric critical care setting confront significant
challenges in the areas of sedation, analgesia, tolerance, and with-
1–2 9 (15.0%) drawal management. The goal of all practitioners is to provide excellent
3–8 19 (31.7%)
care while minimizing patient pain, anxiety, and iatrogenic complica-
tions. To best achieve this balance, clinicians need to fully understand
WAT-1 scores can range from 0 to 12, where a score of 3
or higher is considered to indicate clinically significant sedation management, tolerance and weaning. Pediatric critical care
withdrawal symptoms. nurses play a vital role in this process; their constant bedside presence
provides invaluable perspective and expertise. Providing tools to ac-
curately assess and describe withdrawal symptoms in their patients will
History of Exposure enable nurses to advocate on behalf of this vulnerable patient popula-
tion.

History of Opiate/Benzodiazepine
26.7% Exposure 6.1. Resources to advance clinical inquiry
Opiate/Benzodiazepine Naïve
55.0% At [Boston Children's Hospital], the Nurse Executive Committee for
Research and Inquiry (NECRI), a formal hospital-wide clinical inquiry
18.3% Unknown/Unavailable
infrastructure, supports the advancement of nursing science (JONA
article). Details of this infrastructure have been described previously
(JONA). NECRI, established in 2012, is made up of nurse executives and
PhD prepared nurse scientists that together align clinical inquiry with
nursing clinical operations (JONA). The Nursing Science Fellowship
Fig. 2. History of exposure to opiate/benzos.
(NSF), which is overseen by NECRI, provides each of the participants
with structured mentorship by a PhD prepared nurse scientist to carry
evolve, research needs to include the ever-increasing population of out the completion of a clinical inquiry project. For each of the fellows,
highly complex, chronically ill patients. The potential gap between mentorship starts at the formation of the clinical inquiry question
bedside nurse observation and the WAT-1 tool should be considered through manuscript dissemination. Prior to the start of the NSF, lea-
when determining patient treatment plans. Future work should include dership support of the nursing science fellow is obtained and allotment
the further validation of IWS tools in pediatric patients with multiple of project time is guided by staffing requirements at the unit level. Each
medical conditions. In addition, the role of caregiver input should be of the fellows is encouraged to seek internal and external grant funding
considered, as parents and caregivers know their child best and are opportunities. Details of the NSF can be found in a companion piece
often the first to detect clinical changes. found in this journal issue.

4
P. Conrad, et al. Applied Nursing Research 55 (2020) 151284

Funding 175–183. https://doi.org/10.1097/PCC0000000000000306.

Birchley, G. (2009). Opioid and benzodiazepine withdrawal syndromes in the paediatric
intensive care unit: A review of recent literature. Nursing in Critical Care, 14(1),
This research did not receive any specific grant from funding https://doi.org/10.1111/j.1478-5153.2008.00311.x.
agencies in the public, commercial, or not-for-profit sectors. Fernandez-Carrion, F., Gaboli, M., Gonzalez-Celador, R., Gomez de Quero-Masia, P.,
Fernandez-de Miguel, S., Murga-Herrera, V., ... Payo-Perez, R. (2013). Withdrawal
syndrome in the pediatric intensive care unit. Incidence and risk factors. Medicina
CRediT authorship contribution statement Intensiva, 37, 67–74. https://doi.org/10.1016/j.medinine.2012.02.010.
Fisher, D., Grap, M., Younger, J., Ameringer, S., & Elswick, R. (2013, September 3).
Paula Conrad: Conceptualization, Validation, Formal analysis, Opioid withdrawal signs and symptoms in children: Frequency and determinants.
Heart & Lung, 42, 407–413. https://doi.org/10.1016/j.hrtlng.2013.07.008.
Investigation, Writing - original draft, Writing - review & editing. Franck, L., Scoppettuolo, L., Wypij, D., & Curley, M. (2012). Validity and generalizability
Shannon Meyer: Writing - review & editing. Jon Whiting: of the withdrawal assessment tool-1 (WAT-1) for monitoring iatrogenic withdrawal
Conceptualization, Investigation, Writing - review & editing. Jean syndrome in pediatric patients. Pain, 153, 142–148. https://doi.org/10.1016/j.pain.
2011.10.003.
Anne Connor: Conceptualization, Methodology, Validation, Formal
Ista, E., De Hoog, M., Tibboel, D., Duivenvoorden, H., & Van Dijk, M. (2013, October).
analysis, Investigation, Writing - review & editing, Visualization, Psychometric evaluation of the sophia observation withdrawal symptoms scale in
Supervision. critically ill children. Pediatric Critical Care Medicine, 14(8), 761–768. https://doi.
org/10.1097/PCC.0b)13e31829f5be1.
Ista, E., Van Dijk, M., Gamel, C., Tibboel, D., & De Hoog, M. (2007, June 1). Withdrawal
Declaration of competing interest symptoms in children after long-term administration of sedatives and/or analgesics:
A literature review. “Assessment remains troublesome”. Intensive Care Medicine, 33,
1396–1406. https://doi.org/10.1007/s00134-0070696-x.
None. Madden, K., Burns, M., & Tasker, R. (2017, June 18). Differentiating delerium from se-
dative/hypnotic-related withdrawal syndrome: Lack of specificity in pediatric critical
References care assessment tools. Pediatric Critical Care Medicine, 6, 580–588. https://doi.org/10.
1097/PCC.0000000000001153.
Shapiro, M., Henderson, C., Hutton, N., & Boss, R. (2017, April). Defining pediatric
Amigoni, A., Mondardini, M., Vittadello, I., Zaglia, F., Rossetti, M., Vitale, F., ... Franck, L. chronic critical illness for clinical care, research, and policy. Hospital Pediatrics, 7(4),
(2017, February). Withdrawal assessment tool-1 monitoring in PICU: A multicenter 236–244. https://doi.org/10.1542/hpeds.2016-0107.
study on iatrogenic withdrawal syndrome. Pediatric Critical Care Medicine, 18(2), Tobias, J. (2000). Tolerance, withdrawal, and physical dependency after long-term se-
https://doi.org/10.1097/PCC.0000000000001054. dation and analgesia of children in the pediatric intensive care unit. Critical Care
Anand, K., Willson, D., Berger, J., Harrison, R., Meert, K., Zimmerman, J., ... Nicholson, C. Medicine, 28(6), 2122–2132.
(2010, April 19). Tolerance and withdrawal from prolonged opioid use in critically ill Whelan, K., Heckmann, M., Lincoln, P., & Hamilton, S. (2015). Pediatric withdrawal
children. Pediatrics, 125, e1208–e1225. https://doi.org/10.1542/peds.2009-0489. identification and management. Journal of Pediatric Intensive Care, 4, 73–78. https://
Best, K., Boullata, J., & Curley, M. (2015, February). Risk factors associated with iatro- doi.org/10.1055/s-0035-1556749.
genic opioid and benzodiazepine withdrawal in critically ill pediatric patients: A
systematic review and conceptual model. Pediatric Critical Care Medicine, 16(2),