Preoperative anxiety level is not associated with postoperative negative behavioral changes in premedicated children

Anesthesia preinduction anxiety in children can according to some studies lead to long‐term anxiety and negative behavioral changes (NBC), while other studies have not found this effect. This secondary analysis from a recent premedication trial comparing clonidine and midazolam aimed to test the relation between preoperative anxiety assessed with modified Yale Preoperative Anxiety Scale (mYPAS) and postoperative NBCs assessed with Post Hospital Behavior Questionnaire (PHBQ), regardless of premedication type.


| INTRODUCTION
High levels of preoperative or anesthesia preinduction anxiety in children are undesirable. Preinduction anxiety may influence many aspects of the perioperative period, including the efficacy of anesthetic drugs, anesthetic emergence, the child's general hospital experience, and more. Specifically, some studies have shown that anxiety/ distress at anesthesia induction can lead to increased postoperative pain, long-term anxiety problems and new or regressive negative behavioral changes (NBCs). [1][2][3][4][5][6] This evidence provides motivation for actively attempting to reduce preanesthetic anxiety levels in children.
However, other studies have not demonstrated an association between preanesthetic anxiety levels and new postoperative NBC. [7][8][9] These results bring into question whether levels of preanesthesia anxiety in children are associated with or could predict postoperative negative outcomes. In this secondary analysis of results from a recently published premedication trial in children, 10   The study was conducted according to the Declaration of Helsinki. This report comprises a planned secondary analysis of the database previously used with the primary goal to assess two different premedications for postoperative NBCs. 10 The trial design details have been published elsewhere. 10 A brief overview is presented here. Inclusion criteria were children 2-7 years with ASA physical status I-II undergoing outpatient ear, nose, or throat surgery at a single center, Skellefteå, Sweden. Exclusion criteria included unplanned surgery, not accepting premedication, or parents not understanding the Swedish language.
On the day of surgery, after informed consent was obtained, participants received one of two alternative premedications (randomly allocated), either clonidine 4 μg/kg orally or midazolam 0.5 mg/kg orally, allowing appropriate times from administration to expected maximum effect (60 and 30 min, respectively) before entering the operating room.
The clinicians caring for the patient, the parents, the pre-anesthesia anxiety assessors, and postoperative assessors were blinded to which premedication was given. In the operating room, with parents accompanying their children, the participants were managed by one standard anesthesia protocol for intravenous access and intravenous anesthetic induction, with inhalational anesthetic maintenance. They all had been prepared with a local anesthetic cream (EMLA, Aspen Nordic) on arm/hand in two places 60 min before. One intravenous (iv) cannula was inserted for intravenous induction with fentanyl 1 μg/kg and propofol 3-4 mg/kg. If venous access was difficult or very stressful for the child, mask induction with sevoflurane-nitrous-oxide was performed. Propofol 1 mg/kg iv bolus was given at the start of anesthesia emergence, to try to prevent agitation F I G U R E 1 CONSORT flowchart from original study. 10 T A B L E 1 Demographic and other factors by preoperative anxiety. upon emergence. 11 Postoperative analgesia was also defined by the same protocol, with paracetamol and possibility for ketorolac or morphine administration based on pain scoring as described in the protocol, with paracetamol and ibuprofen allowed after leaving the post-anesthesia unit.

| Outcomes
The main outcome was new postoperative NBC identified using the Post Hospital Behavior Questionnaire (PHBQ). 13,14 The details concerning this outcome have been presented earlier. 10  To guide significance testing, a p-value level <.05 was used.
No specific sample size calculation was performed for this

| DISCUSSION
The main finding in this secondary analysis of results from this trial 10 is that no association between scores indicative of high preanesthetic anxiety and postoperative new negative behaviors was observed in children premedicated with either clonidine or midazolam. This result does not mean that there is no influence from preanesthesia conditions on postoperative recovery, but rather that it could not be demonstrated in this cohort, and that a large effect is unlikely. This observation appears to be independent of type of premedication.
The preoperative anxiety outcome for the primary assessment was categorical, that is, mYPAS positive or negative based on a widely accepted cut-off of 30 points. 12,[19][20][21] The range of mYPAS scores as a continuous factor and new NBCs 1 week after surgery also showed no association, so there was no indication that very high mYPAS scores had a different relationship to the outcome compared with barely positive mYPAS scores.
T A B L E 7 Univariate analysis PHBQ at 1 week positive (>3 NBCs) by mYPAS. The only factor that was found to be associated with new NBCs was age in months. As noted in the primary study, 10 younger age was a significant factor in developing postoperative NBCs, where premedication type was a primary exposure. Younger age was also associated with higher mYPAS score, but again it was not necessarily the participants with high mYPAS score that developed NBCs postoperatively.
The context for this assessment was that premedication should  where there was steady pre-operative anxiety before any contact with healthcare workers would be missed as a factor. Another consideration for the mYPAS anxiety scoring system is that where some children could have more introverted anxiety, this would not be scored the same in the mYPAS instrument compared to children with extroverted expression of anxiety. This would not change the categorical outcome of positive or negative mYPAS since even being silent or not moving will generate a positive mYPAS result. Only playing and responding in a calm way will generate a negative mYPAS result.
Another limitation is the precision of the PHBQ instrument, which relies on parental interpretation and reporting, which is subjective.
The one-week postoperative reporting interval is a compromise between enough distance from the operation and enough recency for optimal parental engagement. The PHBQ instrument itself is much discussed [26][27][28] concerning optimal scoring and analysis routines, and a validated single approach is not yet widely established. The PHBQ outcome definition used here is explained in some depth in the original trial publication. 10 Premedication can be useful for calming an anxious child and to get a smooth anesthesia induction, but this may not affect the development of meaningful postoperative NBCs. Exactly what causes post-NBCs can be multifactorial and differ from child to child. This is one reason why it is important that the entire care during the hospital stay is supportive and of good quality from the child's perspective.
In conclusion, this secondary analysis of prospectively controlled treatments for a cohort of day surgery children premedicated with either clonidine or midazolam finds no relation between preanesthesia anxiety scores and postoperative new negative behaviors. A true relation between this exposure and these outcomes cannot be excluded based on these findings, but if a small exposure effect exists, it is possible that it could have been diluted by other types of exposures to participants in the perioperative period.

AUTHOR CONTRIBUTIONS
Caroline Zickerman contributed to the study design, data collection and analysis and manuscript writing. Camilla Brorsson contributed to the study design, data analysis and manuscript writing. Magnus Hultin contributed to the study design, data analysis and manuscript writing.
Göran Johansson contributed to the study design, data collection and analysis and manuscript writing. Ola Winsö contributed to the study design, data analysis and manuscript writing. Michael Haney contributed to the study design, data collection and analysis and manuscript writing.