13996576, 2021, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/aas.13818 by INASP/HINARI - INDONESIA, Wiley Online Library on [23/11/2022].

See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
| |
Received: 11 November 2020    Revised: 3 March 2021    Accepted: 6 March 2021

DOI: 10.1111/aas.13818

ORIGINAL ARTICLE

Iatrogenic withdrawal syndrome frequently occurs in paediatric

intensive care without algorithm for tapering of analgosedation

Mette Dokken1,2  | Tone Rustøen3,4 | Lien M. Diep5 | Frode E. Fagermoen6 |

Rakel I. Huse1 | Gudny A. Rosland1 | Ingrid Egerod7 | Gunnar K. Bentsen6

1
Division of Emergencies and Critical Care,
Paediatric Intensive Care Section, Oslo Background: Analgesics and sedatives are key elements to reduce physiological and
University Hospital -­Rikshospitalet, Oslo, psychological stress associated with treatment in paediatric intensive care. Prolonged
Norway
2
drug use may induce tolerance and development of iatrogenic withdrawal syndrome
Institute of Clinical Medicine, University of
Oslo, Oslo, Norway (IWS) during the tapering phase. Our primary aim was to describe the prevalence
3
Division of Emergencies and Critical Care, of IWS among critically ill ventilated patients in two Norwegian paediatric intensive
Department of Research and Development,
Oslo University Hospital, Oslo, Norway
care units (PICUs), and secondary to investigate what motivated bedside nurses to
4
Department of Nursing Science, Faculty of administer additional drug doses.
Medicine, Institute of Health and Society, Methods: Mechanically ventilated patients (n = 40) from newborn to eighteen years
University of Oslo, Oslo, Norway
5 of age, with continuous infusions of opioids and benzodiazepines for 5 days or more,
Department of Biostatistic and
Epidemiology, Oslo University Hospital, were included consecutively from May 2016 to June 2018. By using Withdrawal
Oslo, Norway
Assessment Tool-­1 (WAT-­1) twice daily we recorded the prevalence of IWS.
6
Division of Emergencies and Critical
Care, Department of Anesthesiology, Oslo Additionally, we recorded signs and symptoms that led bedside nurses to administra-
University Hospital -­Rikshospitalet, Oslo, tion extra bolus medication.
Norway
7
Results: Peak WAT-­1 score indicated an IWS prevalence of 95% in this selected
Intensive Care Department, University of
Copenhagen, Rigshospitalet, Copenhagen, group. The first days of the tapering phase were most critical for IWS. The most fre-
Denmark
quent symptoms triggering administration of additional bolus doses were agitation/
Correspondence restlessness, and thiopental and propofol were the bolus drugs used most frequently.
Mette Dokken, Paediatric Intensive
Conclusions: IWS affected 95% of the children having received infusions of opioids
Care Section, Oslo University Hospital,
Sognsvannsveien 20, 0372 Oslo, Norway. and benzodiazepines for 5  days or more in PICUs without a tapering protocol for
Email: [email protected],
these drugs. This calls for implementation and testing of such weaning protocols.
[email protected]

Editorial Comment
Sedation over longer periods in paediatric intensive care is often associated with challenges
concerning both drug tolerance and clinical signs of drug withdrawal when changing drugs or
reducing drug doses. This study assesses an instrument for standardized assessment of clinical
signs of withdrawal syndrome during drug weaning in this setting.

1 |  I NTRO D U C TI O N treatment.1 Analgosedation with opioids and benzodiazepines is

provided to ensure an acceptable level of comfort for the patient,
Children receiving intensive care treatment risk complications to alleviate pain and anxiety, and to avoid accidental removal of life-­
related to the severity of their illness and the complexity of the saving tubes and lines. 2 Use of these drugs for more than a few days

This is an open access article under the terms of the Creative Commons Attribution-­NonCommercial License, which permits use, distribution and reproduction
in any medium, provided the original work is properly cited and is not used for commercial purposes.
© 2021 The Authors. Acta Anaesthesiologica Scandinavica published by John Wiley & Sons Ltd on behalf of Acta Anaesthesiologica Scandinavica Foundation

|
928     
wileyonlinelibrary.com/journal/aas Acta Anaesthesiol Scand. 2021;65:928–935.
 |

13996576, 2021, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/aas.13818 by INASP/HINARI - INDONESIA, Wiley Online Library on [23/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DOKKEN et al.       929

may induce tolerance, physical dependence and development of iat- 2.3 | Inclusion and exclusion criteria
3
rogenic withdrawal syndrome (IWS) in the tapering phase.
IWS is defined as an adverse drug response with a pattern of All consecutive patients <18 years requiring mechanical ventilation and
signs and symptoms including central nervous system hyperirritabil- continuous infusion of opioids and/or sedatives for ≥5 days were eligible
ity, gastrointestinal dysfunction and an autonomic dysfunction when for enrolment. Enrolment was not precluded by the use of other seda-
an analgesic or sedative drug is abruptly stopped, sharply decreased tive agents: eg, clonidine, dexmedetomidine, alimemazine, propofol, thi-
or antagonized in a patient who is physically tolerant. 2-­5 Central ner- opental, levomepromazine, chloral hydrate, phenobarbital or ketamine.
vous system affections include irritability, increased wakefulness, Exclusion criteria were severe nervous system impairment that could
tremulousness, hyperactive deep tendon reflexes, clonus, frequent affect assessment of the sedation level. Patients were included only
yawning, sneezing, delirium and hypertonicity.3 Activation of the once, and when the patient's physician decided to initiate tapering. Data
autonomic nervous system may result in tachycardia, hypertension, collection was closed if the patient was transferred to another hospital.
tachypnea, sweating and fever.3 Gastrointestinal dysfunction may
lead to feeding intolerance with vomiting and diarrhoea.3 Many of
the IWS signs and symptoms will overlap with symptoms related to 2.4 | Study population
the child's disease and also with symptoms of delirium.6 Therefore, a
careful assessment using a validated tool is essential to detect IWS.7 We aimed at recruiting 40 consecutive patients when tapering was
Two instruments for assessing IWS in children, the Withdrawal initiated, and to follow these patients in PICU and general ward until
8,9
Assessment Tool-­1 (WAT-­1) and Sophia Observation Withdrawal 3 days after all analgosedation was stopped, however, with a limita-
Symptoms-­scale (SOS)10,11 are recommended in guidelines.12 The tion of maximum 21 days.
frequency of mixed IWS in prospective studies range from 37% to
77% using WAT-­1 and from 18% to 100% using SOS.13
Most researchers using WAT-­1 use Peak WAT-­1 in their analysis 2.5 | Data collection
when assessing IWS.8,14 However, Peak WAT-­1 might not reflect the
degree to which patients are bothered during the tapering phase be- Data collection included patient demographics and clinical char-
cause only one WAT-­1 score from each patient is used in the analy- acteristics. We selected WAT-­1 as our monitoring tool because a
ses. To supplement Peak WAT-­1, we decided to present also the sum Norwegian version was available and seemed convenient to use at
of the elevated WAT-­1 scores (Sum WAT-­1 ≥3) recorded during the the bedside.15 The tool was translated to Norwegian by following an
study period to better describe the IWS burden over time. established protocol for forward and backwards translation in col-
The primary aim was to describe the prevalence and severity of laboration with the originator Linda Franck. The WAT-­1 score is an
IWS among severely ill patients in two Norwegian PICUs, and sec- 11-­item (12-­point) instrument that screens for signs and symptoms
ondary, to investigate the causes that led bedside nurses to adminis- of opioid-­ and benzodiazepine-­related IWS in PICUs.8,9 The total
ter additional bolus doses of sedatives or analgesics. score is a sum of all 11 items where a score ≥3 indicates IWS.8
Sum WAT-­1 ≥3 is calculated by adding all scores ≥3 during the tapering
period. The number of scores included per day was fixed and pre-­defined.
2 |  M ATE R I A L S A N D M E TH O DS We included two scores per day, one recorded between 7 and 10 o`clock
in the morning, and the second between 5 and 7 o`clock in the afternoon.
2.1 | Study design and setting The use of WAT-­1 was not a part of the daily standard of care in
the two units. Accordingly, members of the study group (MD, FEF,
This prospective, observational study was designed to yield baseline data RIH, GAR, GKB) were trained in the use of the tool and conducted
on IWS from two medical–­surgical PICUs with six and three staffed beds WAT-­1 scores twice daily from the first day of tapering and contin-
at Oslo University Hospital, Norway. The PICUs did not have algorithms ued until 72 hours after the last dose of analgesic or sedative was
in place for tapering opioids and benzodiazepines. The health care provid- administered, but limited to 21 days. Unfortunately we were not able
ers did not use any monitoring tool to assess withdrawal signs and symp- to measure inter-­rater-­reliability, but we trained in the use of WAT-­1
toms, just their professional judgement. The physicians decided how and together, and individual. We also discussed how to use the tool with
when to taper when the patients were recovering. The plan is to re-­do the the professionals who translated the WAT-­1 tool into Norwegian,
study once tapering algorithms have been implemented within the units. and with the originator of the tool, Linda Franck.

2.2 | Ethics 2.6 | Detection of symptoms and signs that led to

bolus medication
The Regional Committees for Medical and Health Research Ethics
approved the study (2016/135) and the parents provided written in- A questionnaire for day, evening and night shift staff was developed
formed consent. ClinicalTrials.gov Identifier: NCT02952846. to record signs and symptoms that motivated bedside nurses to
 |

13996576, 2021, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/aas.13818 by INASP/HINARI - INDONESIA, Wiley Online Library on [23/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
930       DOKKEN et al.

administer additional bolus doses of analgesics and/or sedatives. The (30%) had an infectious disease such as bronchiolitis or sepsis.
questionnaire included: sleep disturbance, tremors, seizures, fever, Nineteen patients had a wide range of rare diseases such as biliary
muscle-­contraction, hallucinations, crying, agitation/restlessness, atresia, DiGeorge syndrome, Apert syndrome and severe combined
sneezing, tachycardia, tachypnoea, hypertension, agitation, loose immune deficiency. We categorized the patients into groups based
stools, nausea and vomiting. The nurses picked from the patients pro on the main reason for PICU admission (Table 1).
re nata (PRN) orders, and often they had a choice between opioids,
benzodiazepines and other medication, such as propofol and thio-
penthal. The nurses documented what motivated additional medica- 3.2 | Prevalence of IWS during the tapering phase
tion and recorded the type and dose of medications administered.
We conducted 1175 twice daily WAT-­1 scores on 40 children, and
95% of the children had a Peak WAT-­1 score of ≥3 during the study
2.7 | Statistical analyses period (Figure 1). The median Peak WAT-­1 score was 5 (IQR 3). The
median Sum WAT-­1 ≥3 score was 15.5 (IQR 31). This number indicates
Patient characteristics are reported as medians and interquartile that the majority of the patients suffered from IWS on at least two
ranges (IQRs) and ranges. Frequencies and percentages are given for scoring points, in practice for more than 24  hours. The Spearman's
categorical variables. Sum WAT-­1 ≥3 was used in our analysis to- correlation coefficient between Peak WAT-­1 and Sum WAT-­1 ≥3 was
gether with the Peak WAT-­1 score. We used the closest number in 0.8 (CI 0.65, 0.89). As outlined in Table  2, the correlation between
cases with missing values when we summed the WAT-­1 scores ≥3. explanatory variables and outcome variables was different regarding
For example, if the morning value was missing and the child scored some medication variables when we compared Peak WAT-­1 and Sum
4 in the evening, we set 4 in the morning. For Peak WAT-­1 missing WAT-­1 ≥3 groups.
values are not an issue, only the highest value recorded is used. In the beginning of the tapering phase the patients were most
Distributions of Sum WAT-­1 ≥3, Peak WAT-­1 and patient charac- at risk for developing IWS (Figure 2). Seven patients were readmit-
teristics were explored and checked for normal distribution by using ted to PICU during the tapering phase, one readmission was due to
histograms and quantile-­quantile plots. The correlation between IWS. This patient scored 8 on WAT-­1 on her return. Other adverse
Peak WAT-­1 and Sum WAT-­1 ≥3 was estimated using Spearman events were considered not to be related to IWS because the pa-
correlation analysis. We used the bootstrap method of percentile tients needed additional analgosedation arising from complications
type to estimate the 95% confidence interval which was based on associated with their illness. Seven patients were reintubated due to
1000 replications/bootstraps. Before conducting the analysis of respiratory failure, six had additional surgery, five patients needed
association, Sum WAT-­1 ≥3 and Peak WAT-­1 scores were catego- another general anaesthesia, one patient needed a change of wound
rized into two groups using the medians as the cut-­offs because of
TA B L E 1   Study group characteristics (N = 40)
non-­normality and non-­monotone covariation between drugs, Peak
WAT-­1 and Sum WAT-­1 ≥3. Patients with Sum WAT-­1 ≥3 and Peak Characteristics N (%)
WAT-­1 above the median value was allocated to the moderate/severe Gender
level group. The remaining patients were allocated to the absent/
Male 20 (50)
mild level group. The associations between patient characteristics,
Admission diagnosis
dose of drug infused, and the categorized groups were inspected by
Respiratory insufficiency 19 (47.5)
using Mann-­Whitney-­Wilcoxon tests. The tests were two-­sided, and
Post-­operative monitoring 13 (32.5)
the significance level was set at .05. The analyses were performed
Multiple organ system failure 5 (12.5)
with IBM SPSS Statistics software (v. 25; IBM SPSS).
Acute liver failure 2 (5)
Acute kidney injury 1 (2.5)

3 |   R E S U LT S Median (IQR)

Age (months) 6 (0.16-­21)

3.1 | Characteristics of the children
Weight (kg) 6.65 (3.78-­11.38)
Ventilation (days) 9 (6-­13)
The sample comprised 40 children, 20 girls and 20 boys (Table  1).
PICU (days) 13 (9-­19)
A total of 42 children were assessed for inclusion, but for two,
Hospital stay (days) 32.5 (13-­51.5)
their parents declined participation. The median age was 6 months
(range, 5  days–­9  years). Thirty-­three patients (80%) were younger Analgosedation treatment (days) 18.5 (12.25-­3 0)

than 2  years and nine patients were newborn. The patients had a Analgosedation treatment before weaning (days) 8.5 (6-­11)

wide range of diagnoses: eg, infants with various congenital malfor- Analgosedation weaning (days) 13.5 (8-­21)
mations such as gastroschisis, oesophageal atresia, omphalocele, Abbreviations: IQR, interquartile range (25th–­75th percentiles); N,
diaphragmatic hernia, volvulus and neonatal sepsis. Twelve patients number; PICU, paediatric intensive care unit.
 |

13996576, 2021, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/aas.13818 by INASP/HINARI - INDONESIA, Wiley Online Library on [23/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DOKKEN et al.       931

F I G U R E 1   Peak WAT-­1 for all patients

(N = 40)

dressing, and one patient had a pneumothorax. Data collection was This is not revealed before one actually checks using a monitoring tool.
continued regardless of re-­admission and reasons for this. The pa- When it comes to the WAT-­1 scoring tool, we think that it`s possi-
tients were followed for a maximum of 21 days and 22 patients were ble that the cut-­off score of ≥3 is too low. Patients may be scored
successfully tapered within that timeframe. Fifteen patients needed higher than 2 in a PICU setting even without the presence of IWS. The
more than 21 days and three patients were discharged from the hos- children may have several overlapping symptoms with clinical signs of
pital before the 21 days of data collection was completed. pain, distress and anxiety. Furthermore, loose stools and vomiting may
have causes other than IWS. Fever is also a sign with different other
possible aetiologies, such as infection, overheating in the incubator
3.3 | Symptoms and signs that led to or tucking-­in the child too tightly. These concerns have also been dis-
administration of additional bolus doses cussed by other researchers17 and consequently, Amirnovin et al18 and
Nelson Sanchez-­Pinto et al19 increased the WAT-­1 cut-­off value to ≥4
The most frequent symptoms motivating additional doses of bolus in their research. That said, raising the cut-­off to ≥4 in our material
medications were most frequently agitation/restlessness (34 pa- would have little influence on the observed prevalence (Figure 1). Our
tients) and sleep disturbance (32 patients), and the most frequent IWS prevalence would have been 82.5%.
sign was tachycardia (Figure 3). The bedside nurses’ most frequent
choice of bolus medication was propofol and thiopental (Table 3).
4.2 | Peak WAT-­1 and Sum WAT-­1 ≥3

4 |  D I S CU S S I O N It is our perception that relying only on Peak WAT-­1 when assess-
ing the burden of IWS in a population is sub-­optimal. Peak WAT-­1
4.1 | Prevalence of iatrogenic withdrawal syndrome does not adequately represent a long lasting burden of IWS in an
individual child, because only one single peak value is utilized. We
The main finding in the present study is an IWS prevalence of 95% assume that by using all the elevated WAT-­1 scores ≥3, we get
based on Peak WAT-­1 scores ≥3. Another important finding was that a valuable supplement to the use of Peak WAT-­1 in the analysis.
the most frequent drug used as rescue medication to treat IWS was Using repeated measures, the Sum WAT-­1 ≥3 represents “the area
thiopental and propofol. under the curve” for IWS in each patient, capturing both sever-
IWS in studies with a similar research design has been reported ity and duration. As shown in Table  2, the explanatory variables
to be between 47% and 77%.8,14,16 The use of Peak WAT-­1 ≥3 in the and drug variables differ between Peak WAT-­1 and Sum WAT-­1
present study is in accordance with the recommendation by Franck ≥3 groups. The variables that were significantly associated with a
et al8 who found an IWS rate of 77% using inclusion criteria similar more severe IWS by using Sum WAT-­1 ≥3 were total hospital stay,
to those used here.8 The high prevalence of IWS in our study was un- PICU post-­inclusion stay and duration of weaning (days) (Table 2).
expected. This could be because the group we have studied here, the In this small sample, we therefore suggest that Sum WAT-­1 ≥3 bet-
patients receiving infusions of opioids and benzodiazepines for 5 days ter discriminates between absent/mild IWS and moderate/severe
or more before tapering, accounts for only a very small portion of the IWS compared to Peak WAT-­1. Total hospital stay and number
total PICU population. It is easily done in day to day practice to over- of tapering days are recognized explanatory risk factors and our
look the major problem IWS represents for a small number of patients. findings thus concur with results in earlier research articles.7,8,20
 |

13996576, 2021, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/aas.13818 by INASP/HINARI - INDONESIA, Wiley Online Library on [23/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
932       DOKKEN et al.

TA B L E 2   Association between patient characteristics and drugs with Peak WAT-­1 and Sum WAT-­1 ≥3 groups

Peak WAT-­1 Sum WAT-­1 ≥ 3

≤5 >5 ≤16 >16

Explanatory variable N = 25 N = 15 P-­value N = 21 N = 19 P-­value

Hospital stay, median days 32 (13-­48) 36 (17-­117) 0.192 22 (11-­41) 45 (22-­117) 0.007
(IQR)
PICU stay after inclusion, 4 (3-­6) 8 (4-­12) 0.043 4 (3-­5) 8 (4-­12) 0.017
median days (IQR)
Ventilator and NIV, median 7 (5-­13) 10 (6-­13) 0.192 7 (6-­13) 10 (6-­14) 0.161
days (IQR)
Duration of analgosedation 9 (7-­19) 18 (9-­21) 0.026 8 (7-­14) 21 (15-­21) <0.001
tapering, median days (IQR)
Duration of alpha-­2 agonist, 16 (6-­21) 21 (14-­21) 0.065 10 (5-­18) 21 (18-­21) 0.001
median days (IQR)

Drugs Median (IQR) P-­value Median (IQR) P-­value

Fentanyl dose before weaning 454 (303-­765) 693 (412-­974) 0.112 427 (296-­681) 774 (322-­1031) 0.044
(μg/kg)
Fentanyl total dose (μg/kg) 642 (346-­879) 972 (493-­1932) 0.106 607 (346-­839) 972 (487-­1561) 0.061
Midazolam dose before 7.9 (2.2-­14.5) 14.1 (9.7-­35.9) 0.033 9.5 (3-­28.8) 10.5 (3.9-­18.4) 0.893
weaning (mg/kg)
Midazolam total dose (mg/kg) 10.2 (2.5-­17.2) 23.8 (12.8-­39.2) 0.010 10.3 (3.5-­31.8) 15.7 (10.7-­3 0.2) 0.668
Morphine dose before weaning 0 (0-­2.4) 0 (0-­1.9) 0.740 0 (0-­1.4) 0 (0-­2.4) 0.668
(mg/kg)
Morphine total dose (mg/kg) 2.4 (0.2-­11) 6.8 (0.9-­14.4) 0.422 2.4(0.04-­8.1) 7 (0.9-­14.4) 0.145
Morphine and fentanyla  dose 24.9 (15.5-­42.1) 34.7 (20.6-­57.9) 0.094 21.5 (14.8-­35.7) 41.2 (19.6-­57.9) 0.039
before weaning (mg/kg)
Morphine and fentanyla  total 34.5 (20.1-­59.4) 63.1 (28.4-­98.5) 0.106 33.5 (18.9-­48.7) 63.1 (28.4-­97.2) 0.044
dose (mg/kg)
Dexmedetomidine dose before 4.7 (0-­86) 0 (0-­50) 0.581 31.0 (0-­85.9) 0 (0-­50.2) 0.376
weaning (μg/kg)
Dexmedetomidine total dose 17 (0-­147) 43 (0-­218) 0.699 53.8 (0-­130.6) 32 (0-­218) 0.979
(μg/kg)
Thiopental dose before 17.7 (6.7-­68.7) 51.6 (0-­159.1) 0.761 20.3 (8.7-­75.4) 32.4 (0-­157.7) 0.768
weaning (mg/kg)
Thiopental total dose (mg/kg) 28.7 (7.9-­89) 144.5 (30.4-­213) 0.040 29 (8.7-­105) 89.5 (6-­198.1) 0.320
Clonidine dose before weaning 7 (0-­27) 6 (0-­32) 0.600 0 (0-­19) 17 (0-­53) 0.065
PO (μg/kg)
Clonidine total dose PO (μg/kg) 70 (23-­147) 100 (68-­120) 0.140 53 (16-­91) 107 (86-­214) <0.001

Abbreviations: IQR, interquartile range (25th–­75th percentiles); NIV, noninvasive ventilation; PICU, paediatric intensive care unit; PO, per oral.
a
The doses for fentanyl were converted to morphine equivalents by multiplication by 50. Tested using Mann–­Whitney–­Wilcoxon tests.

However, difference in the number of days on mechanical venti- benzodiazepine dose of 16.0 mg/kg (IQR 8.5-­31.5) had significantly
lation or non-­invasive positive pressure ventilation did not reach more withdrawal symptoms, which is similar to our findings when
significance (Table 2). based on the moderate/severe Peak WAT-­1level group (N  =  15)
(Table 2). Amigoni et al7 found that the only variable that predicted
IWS was the highest administered dose of benzodiazepine. The
4.3 | Risk of benzodiazepine, opioid and WAT-­1 positive group in their sample had a cumulative benzodiaz-
clonidine use epine load of 24  mg/kg (IQR 10-­36.1), also in agreement with our
findings (Table 2).
Several studies have shown that high cumulative doses of benzo- On the other hand, Da Silva et al20 found a low IWS rate (22.6%),
diazepines and opioids are risk factors for developing IWS. Best even though the cumulative doses of midazolam in the IWS group
et al16 found that children exposed to a pre-­weaning cumulative were median 70.4 mg/kg (IQR 41.7-­106.2). The reason for such high
 |

13996576, 2021, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/aas.13818 by INASP/HINARI - INDONESIA, Wiley Online Library on [23/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DOKKEN et al.       933

F I G U R E 2   WAT-­1 ≥ 3 during the tapering phase (N = 40)

F I G U R E 3   Symptoms and signs who led to bolus medication (N = 40)

TA B L E 3   Bolus medication for withdrawal symptoms during the tapering phase (N = 40)

Bolus doses
Number of patients who received no bolus (%) Number of patients who received bolus (%) given(range)

Propofol 15 (37.5) 25 (62.5) 1-­79

Alimemazine 14 (35) 26 (65) 1-­27
Thiopental 21 (52.5) 19 (47.5) 1-­79
Morphine 16 (40) 24 (60) 1-­3 0
Fentanyl 20 (50) 20 (50) 1-­3 0
Midazolam 27 (67.5) 13 (32.5) 4-­54
Ketobemidone 34 (85) 6 (15) 1-­15

doses of midazolam might be that the children exclusively received was exposed to 23.8 mg/kg (IQR 12.8-­39.2) midazolam and if we use
midazolam and fentanyl for sedation in their study. 20 Normally, PICU Sum WAT ≥3 (value >16), the dose was 15.7 mg/kg (IQR 10.7-­30.7)
patients receive different types of analgosedation. 21 In the present (Table 2). High doses of midazolam were significantly associated with
study, the Peak WAT-­1 group scoring >5 (moderate to severe IWS) a more severe IWS when we used the Peak WAT-­1 (P =.010), but was
 |

13996576, 2021, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/aas.13818 by INASP/HINARI - INDONESIA, Wiley Online Library on [23/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
934       DOKKEN et al.

not significant using Sum WAT-­1 ≥3 (P  =.668). This could suggest risk of pulling out invasive equipment. This extensive use of propofol
that midazolam is not the most problematic drug in this small selec- and thiopental may, however, have contributed to the high preva-
tion of patients. This finding would be in agreement with that of Da lence of IWS found in our study (Table 3).
Silva et al, who described low IWS prevalence even with exposure to
very high doses of midazolam. 20
Fentanyl was used as the first line opioid in the present study. 4.5 | Clinical implications
As outlined in Table 2, the dose of fentanyl before weaning was as-
sociated with severe IWS when using the variable Sum WAT-­1 ≥3 IWS may lead to a complex array of signs and symptoms, so doc-
(P =.044). Earlier research has shown that fentanyl increases the risk tors and nurses should regularly assess patients during tapering of
14,22
of developing IWS. The association of IWS with the pre-­t apering analgosedation with an IWS scoring tool, not only rely on individual
dose of fentanyl, but not midazolam in our study, points to fentanyl judgements. A pharmacological treatment algorithm may prevent
as more likely to promote a more long lasting IWS (Table 2). This ob- IWS because then a step-­by-­step guide may contribute strategies
servation is consistent with findings reported by Amigoni et al14 They for prevention and treatment of IWS. In our opinion, there is an ur-
further reported that patients receiving morphine as their primary gent need for testing of weaning algorithms in PICUs. An algorithm
analgesic were 83% less likely to develop IWS than those receiving can provide a common and predictable opioid and benzodiazepine
fentanyl or remifentanil in their multicenter study.14 Our small study tapering regime that does not depend on staff continuity. An algo-
can neither support nor refute these findings as very little morphine rithm needs to be developed that is easy to follow for all healthcare
was given before tapering in our patients (Table 2). It would be very professionals, offering a safe and systematic approach when taper-
relevant to follow-­up this observation in future studies. ing analgosedation in PICU.
In the present study, some children received clonidine during
the entire course, but only to a limited extent before tapering of
analgosedation (Table 2). A high total dose of clonidine was signifi- 5 | LI M ITATI O N S
cantly associated with a more severe IWS (P <.001) (Table 2). Thus,
we believe that clonidine was used in this sample when tapering of Several limitations in the present study need to be addressed. First,
opioids and benzodiazepines was observed to be problematic. In we had a small sample size, the children had a wide range of diagno-
fact, clonidine is widely used to prevent IWS in the weaning phase, ses, and they were treated with different types of analgosedation.
but evidence of its effectiveness is limited. 23 In a recent systematic Second, regarding the use of WAT-­1, the study group (five healthcare
review, dexmedetomidine and clonidine were deemed as options professionals) discussed how to score the children and all positive
for the treatment and prevention of IWS, but the finding did not scores that could cause bias were removed; eg, if the child had loose
24
reach statistical significance. The use of clonidine may lead to a stools due to clostridium or fever due to an infection. However, we
reduction in total doses of midazolam and fentanyl given in PICU. 25 were not able to test the interrater reliability between paired as-
Another drug used in preventing and treating opioid withdrawal is sessment scores. The questionnaire, developed to identify signs or
methadone, and in a paper from 2020, the authors determine that symptoms that led to administration of additional bolus doses of an-
26
this medication can be safely used in PICU, and is recommended algesics or sedatives, was not validated and in some cases the nurses
in guidelines. 27 However, in our PICUs, we have no tradition for the documented several symptoms as the cause of one bolus. This infor-
use of methadone. mation was clarified with the nurses later, which presented a chal-
lenge because some symptoms could not be separated to justify one
additional bolus.
4.4 | Symptoms and signs that led to
administration of extra bolus medicine
6 | CO N C LU S I O N
The most frequent symptoms that led to administration of extra
bolus medicine were agitation/restlessness, sleep disturbance This study demonstrated a very high prevalence of IWS in children
and crying, and the most frequent sign was tachycardia (Figure 3). exposed to long lasting infusions of opioids and benzodiazepines in
Normally the bolus medication, called rescue medicine in IWS treat- PICUs without a written protocol for tapering of analgosedation, and
ment, is an opioid or a benzodiazepine, but in our sample the nurses not using any validated IWS monitoring tool. Additional bolus medi-
frequently administered propofol and thiopental. This is a practice cation administered was primarily not an opioid, and this might have
also reported by others. A web-­based survey from United Kingdom contributed to the high prevalence of IWS. Implementation and test-
showed that propofol (89%) was the most common choice of seda- ing of weaning protocols are needed.
28
tive, followed by midazolam and morphine (49%). Propofol and
thiopental are short-­and rapid-­acting medications, 29 a simple choice AC K N OW L E D G E M E N T S
in a busy everyday PICU, especially if the patients are awake with The authors thank the intensive care nurses and the medical staff of
symptoms as agitation, crying and sleep disturbance, and they pose a the PICUs, and all the healthcare providers involved in the children`s
 |

13996576, 2021, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/aas.13818 by INASP/HINARI - INDONESIA, Wiley Online Library on [23/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DOKKEN et al.       935

ward at Rikshospitalet and Ullevaal, Oslo University Hospital, for col- 14. Amigoni A, Mondardini MC, Vittadello I, et al; Network of Paediatric
Intensive Care Unit Study G. Withdrawal assessment tool-­1 moni-
laboration in the conduct of this study. Thanks to Dr Linda Franck for
toring in PICU: a multicenter study on iatrogenic withdrawal syn-
the permission to use WAT-­1 in this study, and to Astri Maria Lang drome. Pediatr Crit Care Med. 2017;18:e86-­e91.
and Silje Ingvild Fuglseth for the translation of WAT-­1 to Norwegian 15. Suddaby EC, Josephson K. Satisfaction of nurses with the
and permission to use their translation. Withdrawal Assessment Tool-­1 (WAT-­1). Pediatr Nurs. 2013;39:238-­
242, 59.
16. Best KM, Wypij D, Asaro LA, Curley MA. Randomized evaluation
C O N FL I C T O F I N T E R E S T of sedation titration for respiratory failure study I. Patient, process,
The authors have no conflicts of interest. and system predictors of iatrogenic withdrawal syndrome in criti-
cally Ill children. Crit Care Med. 2017;45:e7-­e15.
ORCID 17. Ista E, van Dijk M. Knowing risk factors for iatrogenic withdrawal
syndrome in children may still leave us empty-­handed. Crit Care
Mette Dokken  https://orcid.org/0000-0001-6698-9120
Med. 2017;45:141-­142.
18. Amirnovin R, Sanchez-­Pinto LN, Okuhara C, et al. Implementation
REFERENCES of a risk-­stratified opioid and benzodiazepine weaning protocol in a
1. Motta E, Luglio M, Delgado AF, Carvalho WB. Importance of the pediatric cardiac ICU. Pediatr Crit Care Med. 2018;19:1024-­1032.
use of protocols for the management of analgesia and sedation in 19. Sanchez-­Pinto LN, Nelson LP, Lieu P, et al. Implementation of a risk-­
pediatric intensive care unit. Rev Assoc Med Bras. 2016;62:602-­609. stratified opioid weaning protocol in a pediatric intensive care unit.
2. Anand KJ, Willson DF, Berger J, et al.; Eunice Kennedy Shriver J Crit Care. 2018;43:214-­219.
National Institute of Child H, Human Development Collaborative 20. da Silva PS, Reis ME, Fonseca TS, Fonseca MC. Opioid and benzo-
Pediatric Critical Care Research N. Tolerance and withdrawal from diazepine withdrawal syndrome in PICU patients: which risk factors
prolonged opioid use in critically ill children. Pediatrics. 2010;125:e1 matter? J Addict Med. 2016;10:110-­116.
208-­e1225. 21. Kudchadkar SR, Yaster M, Punjabi NM. Sedation, sleep promotion,
3. Tobias JD. Tolerance, withdrawal, and physical dependency after and delirium screening practices in the care of mechanically venti-
long-­term sedation and analgesia of children in the pediatric inten- lated children: a wake-­up call for the pediatric critical care commu-
sive care unit. Crit Care Med. 2000;28:2122-­2132. nity*. Crit Care Med. 2014;42:1592-­1600.
4. Edwards IR, Aronson JK. Adverse drug reactions: definitions, diag- 22. Fernandez-­C arrion F, Gaboli M, Gonzalez-­Celador R, et al.
nosis, and management. Lancet. 2000;356:1255-­1259. Withdrawal syndrome in the pediatric intensive care unit. Incidence
5. Cunliffe M, McArthur L, Dooley F. Managing sedation withdrawal and risk factors. Med Intensiva. 2013;37:67-­74.
in children who undergo prolonged PICU admission after discharge 23. Duffett M, Koop A, Menon K, Meade MO, Cook DJ. Clonidine for
to the ward. Paediatr Anaesth. 2004;14:293-­298. the sedation of critically ill children: a systematic review. J Pediatr
6. Madden K, Burns MM, Tasker RC. Differentiating delirium from Intensive Care. 2012;1:5-­15.
sedative/hypnotic-­related iatrogenic withdrawal syndrome: lack 24. Ávila-­Alzate JA, Gómez-­Salgado J, Romero-­Martín M, Martínez-­
of specificity in pediatric critical care assessment tools. Pediatr Crit Isasi S, Navarro-­Abal Y, Fernández-­García D. Assessment and treat-
Care Med. 2017;18:580-­588. ment of the withdrawal syndrome in paediatric intensive care units:
7. Amigoni A, Vettore E, Brugnolaro V, et al. High doses of ben- systematic review. Medicine (Baltimore). 2020;99:e18502.
zodiazepine predict analgesic and sedative drug withdrawal 25. Salarian S, Khosravi R, Khanbabaei G, Bagheri B. Impact of oral
syndrome in paediatric intensive care patients. Acta Paediatr. clonidine on duration of opioid and benzodiazepine use in mechan-
2014;103:e538-­e543. ically ventilated children: a randomized, double-­blind, Placebo-­
8. Franck LS, Scoppettuolo LA, Wypij D, Curley MA. Validity and gen- Controlled Study. Iran J Pharm Res. 2019;18:2157-­2162.
eralizability of the Withdrawal Assessment Tool-­1 (WAT-­1) for mon- 26. Friedman SD, Kovach JR, Thompson NE. Methadone's effect on
itoring iatrogenic withdrawal syndrome in pediatric patients. Pain. cardiac repolarization: safety in the PICU. Pediatr Crit Care Med.
2012;153:142-­148. 2020;21:e747-­e751.
9. Franck LS, Harris SK, Soetenga DJ, Amling JK, Curley MA. The 27. Mondardini MC, Vasile B, Amigoni A, et al. Update of recommen-
Withdrawal Assessment Tool-­1 (WAT-­1): an assessment instrument dations for Analgosedation in pediatric intensive care unit. Minerva
for monitoring opioid and benzodiazepine withdrawal symptoms in Anestesiol. 2014;80:1018-­1029.
pediatric patients. Pediatr Crit Care Med. 2008;9:573-­580. 28. Yassin SM, Terblanche M, Yassin J, McKenzie CA. A web-­based sur-
10. Ista E, de Hoog M, Tibboel D, Duivenvoorden HJ, van Dijk M. vey of United Kingdom sedation practice in the intensive care unit.
Psychometric evaluation of the Sophia Observation withdrawal J Crit Care. 2015;30:436.e1:436.e6.
symptoms scale in critically ill children. Pediatr Crit Care Med. 29. Baarslag MA, Allegaert K, Knibbe CA, van Dijk M, Tibboel D.
2013;14:761-­769. Pharmacological sedation management in the paediatric intensive
11. Ista E, van Dijk M, de Hoog M, Tibboel D, Duivenvoorden HJ. care unit. J Pharm Pharmacol. 2017;69:498-­513.
Construction of the Sophia Observation withdrawal Symptoms-­
scale (SOS) for critically ill children. Intensive Care Med.
2009;35:1075-­1081. How to cite this article: Dokken M, Rustøen T, Diep LM, et al.
12. Harris J, Ramelet AS, van Dijk M, et al. Clinical recommendations for Iatrogenic withdrawal syndrome frequently occurs in
pain, sedation, withdrawal and delirium assessment in critically ill
paediatric intensive care without algorithm for tapering of
infants and children: an ESPNIC position statement for healthcare
professionals. Intensive Care Med. 2016;42:972-­986. analgosedation. Acta Anaesthesiol Scand. 2021;65:928–­935.
13. Duceppe MA, Perreault MM, Frenette AJ, et al. Frequency, risk https://doi.org/10.1111/aas.13818
factors and symptomatology of iatrogenic withdrawal from opi-
oids and benzodiazepines in critically Ill neonates, children and
adults: a systematic review of clinical studies. J Clin Pharm Ther.
2019;44:148-­156.