Article Text

PDF

Original research
Effects of music therapy on anxiety and physiologic parameters in angiography: a systematic review and meta-analysis
  1. Adam C Lieber1,
  2. Javin Bose1,
  3. Xiangnan Zhang1,
  4. Hayley Seltzberg1,
  5. Joanne Loewy2,
  6. Andrew Rossetti2,
  7. J Mocco1,
  8. Christopher P Kellner1
  1. 1 Department of Neurosurgery, Mount Sinai Hospital, New York, USA
  2. 2 The Louis Armstrong Center for Music and Medicine, New York, USA
  1. Correspondence to Mr. Adam C Lieber, Department of Neurosurgery, Mount Sinai Hospital, New York, NY 10029, USA; adam.lieber{at}icahn.mssm.edu

Abstract

Background Given the anxiety patients experience during angiography, evidence supporting the efficacy of music therapy during these angiographic procedures is potentially of clinical value.

Objective To analyze the existing literature forthe use of music therapy during cerebral, coronary, and peripheral angiography to determine whether it improves patient anxiety levels, heart rate, and blood pressure during the procedure.

Methods PubMed, Embase, and Scopus were searched to identify studies of interest. Inclusion criteria included studies reporting using music therapy in either cerebral, coronary, or peripheral angiography. Studies focused on a pediatric population; animal studies and case reports were excluded. Participant demographics, interventions, and outcomes were collected by two study authors. Bias and study quality of randomized controlled trials (RCTs) were assessed using the Cochrane Risk of Bias Tool. Separate meta-analyses of the RCTs were performed to compare State Trait Anxiety Inventory (STAI), heart rate (HR), and systolic and diastolic blood pressure (SBP and DBP) in the music intervention group versus control group. Heterogeneity was determined by calculating I2 values, and a random-effects model was used when heterogeneity exceeded 50%.

Results The preprocedure to postprocedure improvement in STAI was significantly greater in the experimental group than the control group (p=0.004), while the decrease in HR, SBP, and DBP was not significant.

Conclusions Recorded music and/or music therapy in angiography significantly decreases patients’ anxiety levels, while it has little to no effect on HR and BP. This meta-analysis is limited by the relatively few RCTs published on this subject.

PROSPERO registration number CRD42018099103

  • angiography
  • blood pressure
  • catheter
  • blood flow

Statistics from Altmetric.com

Introduction

From its beginnings in the early 1900s, when Dr Evan Kane wrote in a JAMA article ‘…I have been employing a phonograph in my operating-room as a means of calming and distracting my patients from the horror of the situation,’ the use of music and/or music therapy has evolved into evidence-based practice in the hospital setting.1

One example of therapeutic music application is in the intensive care unit, which a 2013 JAMA randomized controlled trial (RCT) showed to reduce anxiety of critically ill patients on mechanical-ventilatory support.2 Additionally, in a 2015 meta-analysis in the Lancet, Hole et al found that among 73 studies, music listening improved measures of postoperative pain and decreased the administration of pain medication, improved anxiety, and improved patient satisfaction.3

Music therapy is the use of music and music experiences by a board-certified, licensed music therapist, to deal with the physical, emotional, developmental, and psychological needs of a variety of patient populations. Additional studies have claimed that music helps to relieve self-reported pain, reduce anxiety, and reduce the use of analgesic and anxiolytic drugs in a variety of patient populations, but not all of these studies included true music therapy.3–7

Given the evidence supporting music interventions for surgical and oncology patients, exploring the use of music therapy in coronary, cerebral, and peripheral angiography may advance the scope of its application. During all three types of angiography procedures, different types of catheters are guided through a sheath in the femoral or radial artery into various parts of the body using endovascular pathways. Patients are often awake and receive only local anesthetic at the insertion point of the sheath. This may result in discomfort, pain, and emotional distress as the catheters are being manipulated in the body. Additionally, patients frequently report feelings of heat and burning during the injection of contrast, with certain contrast agents causing more intense sensations than others,8 potentially contributing to exacerbation of anxiety and distress that may negatively influence patient outcomes. This has led to an increase in alternative approaches to improve the angiography experience—for example, recent studies in neurointerventional and cardiology patients have demonstrated the safety, reduced postprocedural pain, and ability to ambulate earlier when performing transradial angiography versus transfemoral angiography.9 10 Music therapists are trained in assessing fear and pain, and apply culturally sensitive musical that is entrained to patients’ vital signs in real time.

In addition to the physical discomfort that some patients report, patients may also experience resultant anxiety about the procedure itself and/or the pathology that the procedure may reveal. In 159 patients undergoing coronary angiography, Gallagher et al found that 37% of patients experienced anxiety related to the diagnostic outcomes of the procedure.11 Tel et al found that for patients undergoing angiographic procedures, factors exacerbating their anxiety included hearing negative stories about friends or family members who had undergone similar procedures and not being introduced to the angiography team before the procedure, indicating that there are psychosocial triggers that can be quelled through interventions.12 This systematic review and meta-analysis surveys the existing literature for the use of music in cerebral, coronary, and peripheral angiography. It provides a platform to further determine if music therapy can bolster improvement in patients’ anxiety levels and physiologic indices of anxiety, such as raised heart rate and blood pressure.

Methods

Data sources and search strategy

This systematic review included studies retrieved from the PubMed, Embase, and Scopus databases. The search and data collection were performed according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.13 A review protocol exists under the PROSPERO registration number: CRD42018099103. Searches were performed on each database to retrieve all literature since 1990 pertaining to the uses of music and music therapy in cerebral angiography, cardiac catheterization, and peripheral angiography. Articles were identified through combinations of various keyword entries including: ‘music therapy,’ ‘music,’ ‘cerebral angiography,’ ‘peripheral angiography,’ and ‘coronary angiography’ (see search strategy in table 1).

Table 1

PubMed search strategy

The literature search is outlined in figure 1. First, a review of titles was performed, in which papers that were irrelevant to topic, non-English language, animal studies, pediatric studies, case reports, and reviews were excluded from analysis. Second, an abstract review was performed, in which papers were excluded for the same reasons as for the title review. Third, a full text review was performed. Finally, the bibliographies of each of the reviews were searched to identify any additional relevant studies that we might have missed. Each step was performed independently by two of the study authors (ACL and JB) and then jointly adjudicated following the reviews.

Figure 1

Search strategy and criteria.

Study selection

Inclusion criteria were any study reporting the use of ‘music therapy’ intervention in cerebral angiography, cardiac catheterization, or peripheral angiography. Exclusion criteria included pediatric studies, animal studies, non-English studies, case reports, and angiography technique reports. Although the distinction between ‘music,’ which can be medicinally informed, and ‘music therapy,’ which involves application of a music therapy assessment and treatment plan adjusted uniquely for the individual patient,14 was considered, studies including either of these interventions were included.

Data extraction and study bias assessment

Angiography type, inclusion criteria, exclusion criteria, number of total patients, mean age, gender breakdown, number of groups, group names, interventions, primary outcome, secondary outcomes, follow-up time period, study limitations, study quality and bias as assessed by the Cochrane Risk of Bias Tool, level of evidence, and discussion of withdrawals and dropouts were collected from each study. The Cochrane Risk of Bias Tool assesses bias by considering selection bias via random sequence generation and allocation concealment, performance bias via blinding of participants and personnel, detection bias via blinding of the outcome assessor, attrition bias via incomplete outcome data, risk of bias across studies via selective reporting within studies, and other bias.15 The results are reported in table 3.

Outcomes collected

The outcomes identified were state trait anxiety inventory (STAI), heart rate (HR), systolic blood pressure (SBP), and diastolic blood pressure (DBP). Studies that reported one or more of these variables in an experimental music therapy group and a control non-music therapy group both before and after the angiography procedure were included in the meta-analysis.

Data analysis

All analyses were performed in RStudio (RStudio 2017, version 3.4.3). Heterogeneity was quantified using τ2 and I2 (with 95% CI). The τ2 statistic was determined using the DerSimonian-Laird estimator. We performed four separate methods to best analyze the variables identified above. For STAI, HR, SBP, and DBP, respectively, first, the pooled means of the studies’ experimental groups before and after the procedure were compared. Second, the pooled means of the studies’ control group between preprocedure and postprocedure were compared. Third, the pooled means of the change in a given parameter between pre- and postprocedure were compared between control and experimental groups. Fourth, the pooled means of postprocedure parameter values were compared between control and experimental groups. Fixed-effects and random-effects models were applied and reported in the forest plots for all meta-analyses performed. For the purpose of data interpretation, a fixed-effects model was treated as the appropriate test for analyses with low heterogeneity (I2<50%) while a random-effects model was used for high heterogeneity (I2>50%). Significance was set at p<0.05.

Results

Study characteristics

Eighteen studies met our ‘systematic review’ criteria based on the fact that they reported some physiological or anxiety-related outcomes in patients receiving music therapy versus patients not receiving music therapy. However, on further review, six of these studies were not included in any of the meta-analyses owing to overlapping of patient populations with another study (n=1), lacking a control group (n=1), or failing to report STAI, SBP, DBP, or HR before and after the intervention or including measures that were incompatible with other studies for meta-analysis, such as abridged STAI scores instead of full STAI scores (n=4). The overall study characteristics of the 12 studies included in the meta-analysis can be found in table 2. The mean number of patients in each study was 90.8 (range 30–200) at baseline and the mean age of patients was 58.5 years (range 42.1–66.8). Where reported, men comprised 74.7% of the study population (range 19–99%). Eight studies (67%) focused on cardiac angiography,16–23 two (17%) on cerebral angiography24 25 and two (17%) on peripheral angiography.26 27

Table 2

Overall study characteristics

Specific characteristics of each study can be found in online supplemental table 1. Ten studies (56%) were eligible for inclusion in the STAI meta-analysis.16 18 19 21–27Eight studies (44%) were eligible for inclusion in the heart rate (HR) meta-analysis.16 17 19–21 23 24 26 Six studies (33%) were eligible for inclusion in the systolic blood pressure (SBP) meta-analysis.16 17 19 20 24 26 Five studies (28%) were eligible for inclusion in the diastolic blood pressure (DBP) meta-analysis.16 17 19 20 26

Quality of studies and risk of study bias

The results of the assessment of each study using the Cochrane Risk of Bias Tool is shown in table 3.

Table 3

Risk of bias

Meta-analysis of collected outcomes

Impact of music therapy on State Trait Anxiety Inventory (STAI)

The pre- to postprocedure decrease in STAI was significantly greater in the experimental group than the control group (figure 2A; mean difference (MD) −2.81, 95% CI −4.72 to −0.91; p=0.004 in the random-effects model, I2=100%). Additionally, the postprocedure STAI showed a trend towards a lower value in the experimental group than the control group (figure 2B; MD −5.08, 95% CI −10.66 to 0.50; p=0.07 in the random-effects model, I2=97%). The decrease in STAI from pre- to postprocedure was found to be significant in the experimental music therapy group alone (online supplemental figure 2C; MD −6.12, 95% CI −8.18 to −4.06; p=0.0001 in the random-effects model, I2=72%) and the control group alone (online supplemental figure 2D; MD −2.99, 95% CI −4.85 to −1.13; p=0.002 in the random-effects model, I2=69%).

Figure 2

The change in preprocedure to postprocedure State Trait Anxiety Inventory (STAI) scores in the music therapy versus control group (A) and the postprocedure STAI score in the music therapy versus the control group (B).

Impact of music and/or music therapy on heart rate

The pre- to postprocedure decrease in HR was not significantly different in the experimental group and control group (figure 3A; MD −0.24, 95% CI −1.47 to 0.98; p=0.70 in the random-effects model, I2=95%). Additionally, the postprocedure HR was not significantly different between experimental and control groups according to the random -effects model (figure 3B; MD −0.91, 95% CI −3.48 to 1.65; p=0.49, I2=54%). The HR decrease from pre- to postprocedure was found to show a trend towards significance in the experimental music therapy group alone (online supplemental figure 3C; MD −3.79, 95% CI −7.82 to 0.24; p=0.07 in the random-effects model, I2=77%) and the control group alone (online supplemental figure 3D; MD −3.28, 95% CI −6.93 to 0.37; p=0.08 in the random-effects model, I2=81%).

Figure 3

The change in preprocedure to postprocedure heart rate (HR) in the music therapy versus control group (A) and the postprocedure HR in the music therapy versus the control group (B).

Impact of music and/or therapy on blood pressure

The pre- to postprocedure decrease in SBP was not significantly different in the experimental group and control groups (figure 4A; MD −5.43, 95% CI −12.74 to 1.88; p=0.15 in the random-effects model, I2=100%). Additionally, the postprocedure SBP was not significantly different between the experimental and control groups (figure 4B; MD −4.66, 95% CI −11.67 to 2.35; p=0.19, I2=81%). The SBP decrease from pre- to postprocedure was found not to be significant in the experimental music therapy group alone (online supplemental figure 4C; MD −12.46, 95% CI −33.06 to 8.14; p=0.24 in the random-effects model, I2=97%) and the control group alone (online supplemental figure 4D; MD −7.05, 95% CI −24.70 to 10.70; p=0.44 in the random-effects model, I2=97%).

Figure 4

The change in preprocedure to postprocedure systolic blood pressure (SBP) in the music therapy versus control group (A) and the postprocedure SBP in the music therapy versus the control group (B).

The pre- to postprocedure decrease in DBP was not significantly different in the experimental and control groups (figure 5A; MD −7.34, 95% CI −16.90 to 2.22; p=0.13 in the random-effects model, I2=100%). Additionally, the postprocedure DBP was not significantly different between the experimental and control groups (figure 5B; MD −4.36, 95% CI −11.38 to 2.66; p=0.22 in the random-effects model, I2=91%). The DBP decrease from pre- to postprocedure was found to be significant in the experimental music therapy group alone (online supplemental figure 5C; MD −2.44; 95% CI −4.25 to −0.64, p=0.01 in the fixed-effects model, I2=0%), but not in the control group alone (online supplemental figure 5D; MD 4.37, 95% CI −1.85 to 10.59; p=0.17 in the random-effects model, I2=90%).

Figure 5

The change in preprocedure to postprocedure diastolic blood pressure (DBP) in the music therapy versus control group (A) and the postprocedure DBP in the music therapy versus the control group (B).

Discussion

Music therapy has been used for decades in the hospital setting, but its use for patients undergoing angiographic procedures is novel, partly owing to the novelty of the procedures themselves. In this meta-analysis we sought to identify all literature on the subject of music therapy during angiography and discern what value it holds for patients undergoing cardiac, cerebral, or peripheral angiography procedures.

Despite considerable heterogeneity among the studies, there is strong evidence suggesting that music therapy reduces anxiety for patients undergoing angiographic procedures compared with its effect in controls. While our results showed that both the music and/or music therapy groups and the control groups had significant reductions in their STAI scores from pre- to postprocedure, the music and/or music therapy group (n=458 patients) had an average 2.81 point further reduction in STAI score compared with the control group (n=448 patients) (95% CI −4.72 to −0.91). When examining the 10 individual RCTs that combined to make this result, nine (90%) found that music therapy was more effective than no therapy for anxiety reduction. A previous meta-analysis of six studies focusing on music therapy specifically in cardiac catheterization procedures similarly found that music therapy led to a 4.0 point reduction in mean STAI scores (95% CI −5.5 to −2.4).28 Of note, two studies which reported an abridged version of the STAI score, the Short STAI, were not included in our STAI meta-analysis.17 29

Music and/or music therapy reduced patients' anxiety levels, which suggest that they can play an important role in the patient’s subjective perception of stress while undergoing angiographic procedures. On the contrary, this meta-analysis suggests that music therapy did not significantly affect these same patients’ physiological manifestations of reduced stress, such as a decrease in HR and blood pressure. For HR, the changes in the music therapy group (n=335) compared with the control group (n=323) from pre- to postprocedure were insignificant in the random-effects model (p=0.49).

One potentially confounding variable is the use of sedatives during angiographic procedures and whether they differed between the music therapy and non-music therapy groups. Higher use of sedatives in music therapy groups, for example, might falsely attribute reduced anxiety in the music therapy groups to that of the music intervention itself. We found that four of the 12 studies included in this meta-analysis reported the use of sedative medications periprocedurally. In all four studies, there were no significant differences in the use of sedatives between the music groups and the non-music groups. Argstatter et al noted that for all 14 drugs administered in total, which included diazepam, there were no intergroup differences in the proportion of the group receiving a given drug or the dose rate (p>0.10 for all medications).20 Vanderboom et al reported that there were no significant differences in the total amount of fentanyl (p=0.78) and midazolam (p=0.54) administered to patients in both groups.24 Bally et al did not measure dose rates, but found that there were no significant differences in patients receiving extra analgesics (p=0.35) or anxiolytic medication (p=0.30) between groups.18 Finally, Mandle et al reported that there were no differences in fentanyl administration between the music and non-music groups; however, in a third group receiving an intervention of ‘relaxing’ sounds, the use of fentanyl was markedly decreased compared with the other two groups (p<0.01).27 Of the eight studies that did not report use of sedation or pain medication periprocedurally, four of them made reference to either baseline sedation use or periprocedural medication use as an untested but potentially confounding variable.17 21 23 26

In the cardiac catheterization studies, the cardiac disease process, medications of the patients, and procedural details were inconsistently reported among the included studies. This is relevant given that these factors could influence patients’ HR or BP. Taylor-Piliae et al reported that the subjects’ cardiac disease processes were acute coronary syndrome (n=23, 51%), angina (n=20, 44%), or arrhythmia (n=2, 4%).23 Rejeh et al reported that the average duration of heart disease in the patient population was 8.93±3.81 months; however, they did not provide a breakdown of the specific disease processes.16 Ripley et al reported that 34% of patients had a previous myocardial infarction, 44% of patients had previous percutaneous coronary intervention, and 19% of patients had had previous coronary artery bypass grafting.17 Argstatter et al reported that 34% of patients had known coronary artery disease and 66% of patients has ‘suspected’ coronary artery disease before catheterization.20 For information on baseline medication details, a relevant point of interest would be any medication that might affect blood pressure or heart rate. Chang et al reported that 33.3% of the included patients used ß-blocker antihypertensive medication at baseline.21 Rejeh et al reported that patients receiving antihypertensive drugs were excluded from the study.16 No other studies provided details on patients’ preprocedural medication. For procedural details, a key distinction to make would be how many patients received percutaneous coronary intervention (angioplasty) during catheterization versus angiography alone. Argstatter et al reported that 65 patients (78%) received only angiography while 17 patients (22%) underwent angiography with percutaneous coronary intervention, with these proportions not differing between study groups.20 Bally et al specified that 91 patients (82%) underwent angiograms while 20 patients (18%) underwent angioplasty.18 None of the other cardiac studies reported how many patients underwent additional percutaneous coronary intervention versus angiography alone.

Some evidence suggests that the type of music used may have a varying effect on HR, with ‘exciting music’ causing increased HR and ‘tranquilizing music’ causing a decreased HR.30 The variability of the music used in the studies included in our meta-analysis may be responsible for a lack of a strong response in HR changes between the control and experimental groups over the course of the angiographic procedure. Significant nuances such as this exemplify the need for the inclusion of a music therapist who could evaluate and advise on the music interventions and prescribe programmes based on evaluation, with sensitivity to patients’ age, cultural background, and musical preference.

The music and/or music therapy intervention had no perceivable effect on patients’ SBP from pre- to postprocedure, while it did significantly decrease patients’ DBP from pre- to postprocedure (p=0.01). However, this periprocedural decrease in DBP was not significantly different from that in the control group (p=0.13). In contrast, a meta-analysis of music and/or music therapy in patients with hypertension showed a significant decrease in SBP was achieved in the music therapy group compared with the non-music therapy group (MD −6.58; 95% CI − 9.38 to −3.79) while no significant decrease in DBP was achieved. However, this meta-analysis is limited as only 90 total patients were included in the analysis.31

While STAI scores were found to be significantly improved in the music therapy group versus the non-music therapy group in our meta-analysis consisting of all three types of angiography procedures, the specific effects of music therapy on cerebral angiography are not as clear. This discrepancy is due to one of the two included cerebral angiography studies, Vanderboom et al concluding that periprocedural music therapy had no effect on the ‘stress response’ (measured by HR and SBP) and anxiety (measured by the STAI score) compared with control. However, further analysis of the study of Vanderboom et al reveals several confounding variables that might account for the lack of response in the music therapy group. First, the study’s control group had a significantly lower STAI  score than the music therapy group at baseline (34.7 vs 40.3, where a lower score indicates less anxiety). Although the control group had a significantly larger pre- to postprocedural reduction in STAI score, the large discrepancy in baseline anxiety levels should be taken into account when interpreting these results. Second, the authors note in the discussion, significant procedural differences in the two groups, noting that ‘those participants in the control group required fewer attempts to gain access to the femoral artery and underwent fewer angiographic runs.”24 The authors do not provide specific data on access attempts and angiographic runs, and the influence of this variable was not removed when performing their analysis on STAI score decrease. It is therefore possible that the effect of music therapy was confounded by a higher complexity and longer procedure in the music therapy group. Finally, other differences between the groups included whether angiography was performed before or after the treatment for the patient’s aneurysm or arteriovenous malformation and which nurses were involved in the procedure. It is conceivable that these differences might account for differing experiences between the groups irrespective of the music intervention.24

No longitudinal outcomes were collected in any of the studies analyzed, as none of the studies included any follow-up once the patients left the hospital. Although the basis for music intervention was often to provide momentary comfort and reduce anxiety in patients undergoing invasive procedures, it remains to be seen whether any of the interventions had any long-term effects on anxiety levels. A future study might consider follow-ups at 3, 6, and 12 months to measure these effects, on patients’ reactions to medically induced emotional trauma, which can threaten future compliance and trust of treatment protocols.

Study limitations

Our study has several limitations. First, while a broad literature search was performed across four medical databases, it is possible that there are additional eligible articles that were not included. Second, while a random-effects model was used because of the high heterogeneity in the studies, lower heterogeneity values (I2 <50%) would have been ideal. Third, with respect to the quality of the studies included in the meta-analysis, many studies failed to report randomization techniques, indicating some levels of bias in those studies. Although these factors are not taken into account when weighing studies in a random-effects meta-analysis, we have made efforts to provide a graphical representation of the quality of the respective studies in table 2. Finally, although patient experiences are similar across the various types of angiography procedures surveyed, there are certainly differences in patient populations and procedural details that would have an impact on the effect of music therapy.

Conclusions

The literature on music and music therapy use during angiography, especially cerebral and peripheral angiography, is limited. However, our meta-analysis shows that recorded music in angiography significantly decreases patients’ anxiety levels, while it has little to no effect on physiological parameters such as HR and BP. Scant literature exists on the use of live music therapy protocols in angiography suites to alleviate patient anxiety, paving the way for high-quality randomized controlled trials to study these effects.

References

View Abstract

Footnotes

  • Contributors The authors have all contributed significantly to the composition of the manuscript and critically edited the final version for content. All have agreed to the presented order of authorship.

  • Funding This research was supported in part by a grant from Arminio and Lucyna Fraga and by a grant from Mr and Mrs Durkovic.

  • Competing interests CPK: (1) director of CME course titled ’Endoscopic minimally invasive intracerebral hemorrhage evacuation' funded by Penumbra and (2) recipient of a Siemens Foundation research grant. JB: son of Arani Bose, MD, Chief Medical Officer and Chairman of Penumbra, Inc. JM: National/International PI/Co-PI: INVEST (Co-PI), COMPASS (Co-PI), THERAPY (PI), FEAT (PI), POSITIVE (Co-PI), BARREL (PI); consultant: TSP, Rebound Therapeutics, Viseon, Pulsar, Cerebrotech, Endostream, Vastrax; Investor/Stockholder/Owner: Cerebrotech, TSP, Endostream, Apama, Rebound, Viseon, Neurvana, Cardinal Consulting, BlinkTBI, Serenity, NTI.

  • Patient consent Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Presented at This work has not been previously presented at a conference nor published as a conference abstract.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.