SPINE Volume 39, Number 3, pp E174-E181 ©2014, Lippincott Williams & Wilkins
DEFORMITY
Predictors of Postoperative Pain Trajectories in Adolescent Idiopathic Scoliosis Mark Connelly, PhD,*† R. Dylan Fulmer, BA,*‡ Jennifer Prohaska, PhD,*§ Lynn Anson, RN-BC,* Lisa Dryer, MSN, RN, CPN,* Valorie Thomas, MSN, RN, CS, PNP,* Jill E. Ariagno, MSN, RN, CPNP,* Nigel Price, MD,*† and Richard Schwend, MD*†
Study Design. Prospective observational study. Objective. To determine the typical trajectory of pain during the first 6 months after spinal fusion surgery for adolescent idiopathic scoliosis (AIS) and the extent to which certain demographic, medical, and psychological variables modify this trajectory. Summary of Background Data. Pain after spinal fusion surgery for AIS may not improve predictably with elapsed healing time, and limited data exist on predictors of the course of pain during the initial months after surgery. Methods. Fifty patients ages 11 to 17 (mean = 14.5, standard deviation = 1.9) with AIS and undergoing posterior spinal fusion surgery comprised the study sample. Pain outcomes were assessed at 4 time points after surgery (2-week, 6-week, 3-month, and 6-month postsurgery). Preoperative predictor variables comprising demographics, baseline Cobb angle, body mass index, baseline pain, and psychological variables (anxiety, negative mood, and confidence in ability to control pain) were assessed 2 weeks before surgery. Perioperative predictor variables comprising pain, pain coping efficacy, negative mood, surgery length, length and lowest level of fusion, and analgesic use were assessed by self-report or record review. Multilevel growth models were used to evaluate hypotheses pertaining to predictors of pain trajectories. Results. Pain level on average declined predictably with days since surgery (b = −0.14 to −0.19, P < 0.01). For 22% of adolescents, pain was at or above baseline levels through 6 months after surgery. Greater baseline pain and anxiety predicted slower improvement in pain, whereas greater confidence in ability to control pain predicted
From the *Children’s Mercy Hospitals and Clinics, Kansas City, MO; †University of Missouri-Kansas City School of Medicine, Kansas City, MO; ‡University of Kansas, Lawrence, KS; and §University of Kansas Medical Center, Kansas City, MO. Acknowledgment date: May 14, 2013. First revision date: September 12, 2013. Second revision date: October 15, 2013. Acceptance date: October 21, 2013. The manuscript submitted does not contain information about medical device(s)/drug(s). Cross Foundation intramural award funds were received, in part, to support this work. No relevant financial activities outside the submitted work. Address correspondence and reprint requests to Mark Connelly, PhD, Integrative Pain Management, Children’s Mercy Hospitals and Clinics, 2401 Gillham Rd, Kansas City, MO 64108; E-mail: [email protected] DOI: 10.1097/BRS.0000000000000099
E174
www.spinejournal.com
more rapid declines in pain. None of the demographic or medical variables reliably modified postsurgical pain trajectories. Conclusion. Although pain typically declines predictably with healing time from spinal fusion surgery for AIS, higher preoperative levels of pain and anxiety may be risk factors for chronic postsurgical pain whereas greater pain coping efficacy may help optimize postsurgical pain outcomes. Key words: adolescent, pediatric, scoliosis, spinal fusion, surgery, pain. Level of Evidence: 3 Spine 2014;39:E174–E181
S
pinal fusion surgery for correction of adolescent idiopathic scoliosis (AIS) is one of the most invasive procedures performed in pediatrics and typically results in moderate to high levels of pain.1,2 In many cases the experience of pain is relatively transient and declines predictably over time with recovery from surgery.3 In other cases, however, pain can persist at or beyond baseline severity levels for months to years after spinal fusion and engender ongoing problems in everyday functioning.4,5 Identifying variables that influence the course of pain during the first several months after fusion surgery for AIS therefore has important clinical implications. Only in the past several years has pain been routinely incorporated into the outcome literature on pediatric spinal fusion surgery.4 Little remains known about the trajectories and predictors of pain during the first few months after surgery, a time interval in which the transition from acute to chronic postsurgical pain is most likely to develop.6,7 The extent to which the course of pain after spinal fusion surgery for AIS is a function of preoperative or immediate postoperative variables also is largely unexplored. Additionally, few studies have examined relative contributions of demographic, medical, and psychological variables to predicting postoperative pain in this patient population. In this study, we sought to (1) determine the typical trajectory of pain through 6 months after spinal fusion surgery in patients with AIS; and (2) determine if demographic, medical, and/or psychological variables were reliably predictive of individual differences in postoperative pain trajectories. We hypothesized that on average pain intensity will linearly February 2014
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE130580.indd E174
30/12/13 2:53 PM
DEFORMITY decline through 6 months after spinal fusion surgery, but that adolescents with higher baseline pain, higher baseline anxiety and negative mood, and lower baseline confidence in ability to cope with pain will have poorer improvement in pain during this time interval. Inferring from existing literature, we did not expect demographics (age, sex, socioeconomic status) or medical variables (baseline Cobb angle, time since diagnosis, body mass index, length of spinal fusion, lowest instrumented vertebrae, length of surgery, and amount of postoperative analgesic use) to be significant moderators of postoperative pain trajectories.
Predictors of Pain in AIS • Connelly et al
MATERIALS AND METHODS
breakthrough pain on postoperative days 1 to 3. Children were transitioned to scheduled oral OxyContin (Purdue Pharma L.P., Stamford, CT) and as needed oxycodone/acetaminophen on postoperative day 4 or whenever they were tolerating a regular diet. Children were weaned from oral opioid analgesics typically around 2 weeks posthospitalization. After hospital discharge, children and their primary caregivers completed outcome measures at approximately 2-week, 6-week, 3-month, and 6-month posthospitalization study visits. Children were paid $25 per completed assessment. Visit completion rates were as follows: 98% for 2-week follow-up, 84% for 6-week follow-up, 88% for 3-month follow-up, and 80% for 6-month follow-up.
Participants
Measures
Study participants were recruited during a period of 18 months from an orthopedic clinic at a pediatric academic medical center in the Midwest. Eligibility criteria were as follows: (1) diagnosis of AIS; (2) scheduled for posterior spinal fusion surgery; (3) able to speak and read English; (4) no prior orthopedic surgery; and (5) no comorbid chronic medical condition or significant learning problems. Of the 64 eligible patients approached for participation, 9 declined to participate due to time concerns or lack of interest. Two enrolled patients had their surgery canceled and 3 patients did not show up for their baseline study visit. The final usable sample therefore comprised 50 participants, ranging in age from 11 to 17 years (mean = 14.5, standard deviation = 1.9). The majority (82%) of the sample was female. Most participants identified as Caucasian (87%) and the remainder identified as African American (13%). Median annual combined family income for the sample was $60,000. Participants in the sample on average had been diagnosed with scoliosis 17 months (±15.4 mo) prior to surgery.
Procedure The study procedures were approved by the institutional review board of the hospital. Families were informed about the study by a research assistant during a preoperative visit, and written child assent, and parental permission was obtained from those interested in participating. Children and their primary caregivers then independently completed baseline study measures during a scheduled study visit occurring approximately 2 weeks prior to surgery (range, 1–4 wk). After the child’s surgery, a research assistant administered the perioperative (inpatient) self-report measures twice daily (between 7 AM and 10 AM and between 4 PM and 7 PM) until the child’s hospital discharge. Children received standard perioperative and postoperative pain management per institution protocol. Specifically, children received intrathecal morphine (15 μg/kg) intraoperatively. Postoperatively, children received continuous infusion of hydromorphone (4 µg/kg/hr), or morphine 0.2 mg/kg if hydromorphone was not tolerated, with a patient-controlled analgesia dose of 4 µg/kg and 6-minute lockout and bolus dose of 10 µg/kg for pain ratings above 5/10 up to once every 2 hours. Intravenous ketorolac boluses (0.5 mg/kg) up to once every 6 hours also were used for Spine
Pain Intensity Data on typical and highest back pain intensity were collected by visual analogue scales prior to surgery (as baseline predictors of postoperative pain outcomes), twice daily throughout the postoperative hospital stay until discharge (as perioperative predictors of subsequent pain outcomes), and at each of the 4 follow-up assessments. Children were instructed to mark anywhere along a 100-mm horizontal line with anchors “no pain” to “most pain” to indicate their typical and highest back pain intensity during the past month (at baseline and follow-up) or during the past 4 hours (during the postoperative hospital stay). Intensity scores were derived by measuring the distance in millimeters between the left anchor and where the child’s mark crossed the horizontal line; possible scores were in the range from 0 to 100, with higher scores indicating greater pain intensity. The visual analogue scale is the recommended measure of pain intensity for research with adolescents and is considered to be empirically well established.8,9 Anxiety Level of preoperative anxiety was measured at the baseline visit using the state portion of the State-Trait Anxiety Inventory for Children. The questionnaire contains 20 items for which children are asked to select adjectives that vary in intensity (e.g., “very calm,” “calm,” “not calm”) to best describe how they feel “right now.” The State-Trait Anxiety Inventory for Children is frequently used as an index of preoperative anxiety in children and has extensive support for its validity and internal consistency.10,11 Item scores were reverse coded where applicable and summed to comprise a total score, with higher scores indicating greater baseline anxiety. Negative Mood Mood was measured at all study time points (at baseline, twice daily throughout the hospital stay until discharge, and at each of the 4 follow-up assessments) using a 7-item semantic differential scale.12 Participants were presented with 7 items that varied in mood intensity along a 7-point scale (e.g., “extremely happy” to “extremely unhappy”) and were asked to select a response that best represented how they felt. Item www.spinejournal.com
E175
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE130580.indd E175
30/12/13 2:53 PM
DEFORMITY scores were summed to form a total score, with higher scores indicating greater negative mood. Pain Coping Efficacy Children’s confidence in ability to control pain was assessed using the following question: “Based on all the things you have done to deal with your pain, how confident are you that you could decrease your pain if you tried?”13 Children recorded their response on a 100-mm visual analogue scale with anchors “no confidence” to “complete confidence;” possible scores were in the range from 0 to 100 with higher scores indicating greater confidence in ability to control pain. Demographic Information Information about the age, sex, and race of the child and income level of the family (as a proxy for socioeconomic status) was obtained at baseline by a parent-report questionnaire. Medical Information Information on time since initial diagnosis of AIS, severity of spinal curvature (Cobb angle) at baseline, body mass index, length of fusion (in levels), lowest instrumented vertebrae, and duration of spinal fusion surgery was obtained through medical record review. Analgesic Use Data on daily (24-hour) cumulative opioid analgesic use during the postoperative hospital stay were documented by a research nurse and then averaged across hospital days for use in analyses. The medication data were converted to parenteral morphine-equivalent milligram dosages divided by weight using standard conversion ratios14 to account for multiple drug types and modes of delivery. Whether or not opioids continued to be used at the time of follow-up visits also was recorded.
Analyses Data on study measures were summarized for descriptive purposes using means, variances, and/or frequency distributions. Multilevel growth models with maximum likelihood estimation15 were then specified using HLM software (Scientific Software International Inc., Skokie, IL)16 to test the primary research questions. These models estimate changes over time in a given outcome variable (in this case pain intensity) for each individual and then estimate the overall average change (the intercept coefficient) and the extent to which specified predictor variables shift the average change trajectory (slope coefficients). An unconditional growth model (containing only “time” as a predictor variable) was used to derive an estimate of the typical daily rate of change in typical and highest pain intensity since surgery. Separate conditional growth models were then used to evaluate individual differences in the daily rate of change in pain intensity during the study time period as a function of demographic variables (age, sex, income level), baseline predictor variables (pain level, E176
www.spinejournal.com
Predictors of Pain in AIS • Connelly et al
Cobb angle, body mass index, time since diagnosis, anxiety, negative mood, and pain coping efficacy), and perioperative predictor variables (length of surgery, length of fusion, lowest instrumented vertebrae, postoperative pain levels, analgesic use, negative mood, and pain coping efficacy). Opioid analgesic use at the time of each study visit was entered into models as a covariate to control for the expected association between continued postoperative pain and continued use of analgesics. In addition, the multilevel growth models were specified to evaluate associations between the baseline predictor variables and preoperative pain; this enabled a determination of whether predictors of postoperative pain trajectories were different from variables already associated with pain at baseline.
RESULTS Descriptives Table 1 presents descriptive statistics for the mean pain intensity scores by time point. Almost all children (89%) reported at least mild pain intensity prior to surgery. Typical and highest pain intensity were on average mild to moderate at baseline (mean = 36/100 ± 25.51, and 50/100, respectively), moderate to severe during the hospital stay (mean = 49/100 and 64/100, respectively) and through 2-week follow-up (mean = 46/100 and 65/100, respectively), and then dropped to mild for the remainder of the follow-up periods (mean = 12–20/100 and mean = 23–38/100, respectively). Approximately 1 in 5 children (22%) reported greater pain intensity levels at their final follow-up assessment relative to their baseline (preoperative) levels. Two patients (4%) were on analgesics preoperatively. Most patients (70%) discontinued use of opioid analgesics by the 6-week postoperative follow-up.
Typical Trajectory of Postoperative Pain Figure 1 shows the mean typical and highest pain intensity from baseline through the follow-up period. Both typical and highest pain intensity on average gradually and significantly improved over days since surgery (b = −0.14 ± 0.02, t (49) = 12.33, P < 0.01, and b = −0.19 ± 0.03, t (49) = 6.78, P < 0.01, respectively).
Baseline Predictors of Preoperative Pain None of the demographic, medical, or psychological variables were found through multilevel model analyses to be significantly associated with preoperative typical pain intensity. However, baseline pain coping efficacy was inversely associated with highest pain intensity reported at baseline, b = −0.22 ± 0.10, t (46) = 2.32, P = 0.02.
Baseline Predictors of Postoperative Pain Trajectories Table 2 presents the results of the hierarchical linear models comprising baseline predictors of postoperative pain trajectories. Pain score trajectories did not reliably differ as a function of baseline Cobb angle, body mass index, time since diagnosis, or demographic variables. Children reporting higher baseline pain had a significantly slower daily rate February 2014
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE130580.indd E176
30/12/13 2:53 PM
DEFORMITY
Predictors of Pain in AIS • Connelly et al
TABLE 1. Pain Intensity Descriptive Statistics by Study Time Point Time Point Baseline
Hospital stay*
2-week follow-up
6-week follow-up
3-month follow-up
6-month follow-up
Measure
Mean ± SD
Typical pain level
36.10 ± 24.80
0–98
Highest pain level
49.58 ± 30.39
0–99
Typical pain level
49.47 ± 14.74
18–81
Highest pain level
63.99 ± 15.57
28–90
Typical pain level
45.56 ± 23.23
3–86
Highest pain level
65.29 ± 27.07
8–100
Typical pain level
20.12 ± 20.62
0–78
Highest pain level
37.76 ± 29.59
0–100
Typical pain level
13.91 ± 18.85
0–70
Highest pain level
24.53 ± 28.21
0–83
Typical pain level
11.57 ± 17.56
0–82
Highest pain level
23.48 ± 29.28
0–100
Range
*Obtained on the basis of twice daily assessment averaged over the length of the hospital stay after spinal fusion surgery. SD denotes standard deviation.
of improvement in typical and highest pain intensity, b = 0.002 ± 0.001, t (123) = 3.68, P < 0.01, and b = 0.001 ± 0.001, t (123) = 2.14, P = 0.03, respectively. Children with greater pain coping efficacy at baseline had a significantly more rapid improvement in postoperative typical and highest pain intensity, b = −0.002 ± 0.001, t (123) = −2.64, P < 0.01, and b = −0.002 ± 0.001, t (123) = −2.13, P = 0.04. In addition, children with higher baseline anxiety, but not children with higher baseline negative mood, had a significantly slower daily rate of improvement in postoperative typical and highest pain, b = 0.01 ± 0.003, t (122) = 2.37, P = 0.02, and b = 0.001 ± 0.004, t (122) = 2.37, P = 0.05. Figure 2 and Figure 3 display postoperative pain trajectories as a function of highest and lowest quartile baseline anxiety and pain scores, respectively.
Figure 1. Pain intensity ratings by study time point. FU indicates follow-up. Spine
Perioperative Predictors of Postoperative Pain Trajectories Table 3 presents the results of the hierarchical linear growth models comprising the perioperative predictors of postoperative pain trajectories. Only pain coping efficacy reported during the hospital stay predicted more rapid improvement in typical and highest pain trajectories, b = −0.002 ± 0.001, t (123) = −2.18, P = 0.03, and b = −0.002 ± 0.001, t (123) = −2.29, P = 0.02. Length of surgery, length of fusion, lowest instrumented vertebrae, amount of opioid analgesic use during the hospital stay, and mood ratings during the hospital stay all were unrelated to postoperative pain trajectories.
DISCUSSION This study sought to determine the extent to which certain demographic, biomedical, and psychological variables measured at baseline and/or immediately after spinal fusion surgery for AIS predict differences in the course of adolescents’ pain during the initial 6-month period after the procedure. Pain on average gradually and significantly declined throughout the time interval studied. However, the rate of decline in pain levels was significantly slower for adolescents who prior to surgery reported high levels of pain and anxiety, and approximately 1 in 5 adolescents reported greater pain intensity levels throughout their follow-up assessment compared with their preoperative pain level. Children who prior to surgery and throughout the hospital stay had greater confidence in their ability to cope with pain had lower pain intensity preoperatively and had more rapid improvement in pain levels through 6 months after surgery. Demographic and biomedical variables did not reliably predict differences in pain trajectories. www.spinejournal.com
E177
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE130580.indd E177
30/12/13 2:53 PM
DEFORMITY
Predictors of Pain in AIS • Connelly et al
TABLE 2. Results of Linear Growth Modeling Predicting Daily Changes in Pain Since Surgery From
Baseline Predictors
For Changes in Typical Pain Predictor
b (±SE)
Intercept*
−0.14 (±0.02)
Sex
T
For Changes in Highest Pain P
b (±SE)
P
−7.60
DEFORMITY
Predictors of Postoperative Pain Trajectories in Adolescent Idiopathic Scoliosis Mark Connelly, PhD,*† R. Dylan Fulmer, BA,*‡ Jennifer Prohaska, PhD,*§ Lynn Anson, RN-BC,* Lisa Dryer, MSN, RN, CPN,* Valorie Thomas, MSN, RN, CS, PNP,* Jill E. Ariagno, MSN, RN, CPNP,* Nigel Price, MD,*† and Richard Schwend, MD*†
Study Design. Prospective observational study. Objective. To determine the typical trajectory of pain during the first 6 months after spinal fusion surgery for adolescent idiopathic scoliosis (AIS) and the extent to which certain demographic, medical, and psychological variables modify this trajectory. Summary of Background Data. Pain after spinal fusion surgery for AIS may not improve predictably with elapsed healing time, and limited data exist on predictors of the course of pain during the initial months after surgery. Methods. Fifty patients ages 11 to 17 (mean = 14.5, standard deviation = 1.9) with AIS and undergoing posterior spinal fusion surgery comprised the study sample. Pain outcomes were assessed at 4 time points after surgery (2-week, 6-week, 3-month, and 6-month postsurgery). Preoperative predictor variables comprising demographics, baseline Cobb angle, body mass index, baseline pain, and psychological variables (anxiety, negative mood, and confidence in ability to control pain) were assessed 2 weeks before surgery. Perioperative predictor variables comprising pain, pain coping efficacy, negative mood, surgery length, length and lowest level of fusion, and analgesic use were assessed by self-report or record review. Multilevel growth models were used to evaluate hypotheses pertaining to predictors of pain trajectories. Results. Pain level on average declined predictably with days since surgery (b = −0.14 to −0.19, P < 0.01). For 22% of adolescents, pain was at or above baseline levels through 6 months after surgery. Greater baseline pain and anxiety predicted slower improvement in pain, whereas greater confidence in ability to control pain predicted
From the *Children’s Mercy Hospitals and Clinics, Kansas City, MO; †University of Missouri-Kansas City School of Medicine, Kansas City, MO; ‡University of Kansas, Lawrence, KS; and §University of Kansas Medical Center, Kansas City, MO. Acknowledgment date: May 14, 2013. First revision date: September 12, 2013. Second revision date: October 15, 2013. Acceptance date: October 21, 2013. The manuscript submitted does not contain information about medical device(s)/drug(s). Cross Foundation intramural award funds were received, in part, to support this work. No relevant financial activities outside the submitted work. Address correspondence and reprint requests to Mark Connelly, PhD, Integrative Pain Management, Children’s Mercy Hospitals and Clinics, 2401 Gillham Rd, Kansas City, MO 64108; E-mail: [email protected] DOI: 10.1097/BRS.0000000000000099
E174
www.spinejournal.com
more rapid declines in pain. None of the demographic or medical variables reliably modified postsurgical pain trajectories. Conclusion. Although pain typically declines predictably with healing time from spinal fusion surgery for AIS, higher preoperative levels of pain and anxiety may be risk factors for chronic postsurgical pain whereas greater pain coping efficacy may help optimize postsurgical pain outcomes. Key words: adolescent, pediatric, scoliosis, spinal fusion, surgery, pain. Level of Evidence: 3 Spine 2014;39:E174–E181
S
pinal fusion surgery for correction of adolescent idiopathic scoliosis (AIS) is one of the most invasive procedures performed in pediatrics and typically results in moderate to high levels of pain.1,2 In many cases the experience of pain is relatively transient and declines predictably over time with recovery from surgery.3 In other cases, however, pain can persist at or beyond baseline severity levels for months to years after spinal fusion and engender ongoing problems in everyday functioning.4,5 Identifying variables that influence the course of pain during the first several months after fusion surgery for AIS therefore has important clinical implications. Only in the past several years has pain been routinely incorporated into the outcome literature on pediatric spinal fusion surgery.4 Little remains known about the trajectories and predictors of pain during the first few months after surgery, a time interval in which the transition from acute to chronic postsurgical pain is most likely to develop.6,7 The extent to which the course of pain after spinal fusion surgery for AIS is a function of preoperative or immediate postoperative variables also is largely unexplored. Additionally, few studies have examined relative contributions of demographic, medical, and psychological variables to predicting postoperative pain in this patient population. In this study, we sought to (1) determine the typical trajectory of pain through 6 months after spinal fusion surgery in patients with AIS; and (2) determine if demographic, medical, and/or psychological variables were reliably predictive of individual differences in postoperative pain trajectories. We hypothesized that on average pain intensity will linearly February 2014
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE130580.indd E174
30/12/13 2:53 PM
DEFORMITY decline through 6 months after spinal fusion surgery, but that adolescents with higher baseline pain, higher baseline anxiety and negative mood, and lower baseline confidence in ability to cope with pain will have poorer improvement in pain during this time interval. Inferring from existing literature, we did not expect demographics (age, sex, socioeconomic status) or medical variables (baseline Cobb angle, time since diagnosis, body mass index, length of spinal fusion, lowest instrumented vertebrae, length of surgery, and amount of postoperative analgesic use) to be significant moderators of postoperative pain trajectories.
Predictors of Pain in AIS • Connelly et al
MATERIALS AND METHODS
breakthrough pain on postoperative days 1 to 3. Children were transitioned to scheduled oral OxyContin (Purdue Pharma L.P., Stamford, CT) and as needed oxycodone/acetaminophen on postoperative day 4 or whenever they were tolerating a regular diet. Children were weaned from oral opioid analgesics typically around 2 weeks posthospitalization. After hospital discharge, children and their primary caregivers completed outcome measures at approximately 2-week, 6-week, 3-month, and 6-month posthospitalization study visits. Children were paid $25 per completed assessment. Visit completion rates were as follows: 98% for 2-week follow-up, 84% for 6-week follow-up, 88% for 3-month follow-up, and 80% for 6-month follow-up.
Participants
Measures
Study participants were recruited during a period of 18 months from an orthopedic clinic at a pediatric academic medical center in the Midwest. Eligibility criteria were as follows: (1) diagnosis of AIS; (2) scheduled for posterior spinal fusion surgery; (3) able to speak and read English; (4) no prior orthopedic surgery; and (5) no comorbid chronic medical condition or significant learning problems. Of the 64 eligible patients approached for participation, 9 declined to participate due to time concerns or lack of interest. Two enrolled patients had their surgery canceled and 3 patients did not show up for their baseline study visit. The final usable sample therefore comprised 50 participants, ranging in age from 11 to 17 years (mean = 14.5, standard deviation = 1.9). The majority (82%) of the sample was female. Most participants identified as Caucasian (87%) and the remainder identified as African American (13%). Median annual combined family income for the sample was $60,000. Participants in the sample on average had been diagnosed with scoliosis 17 months (±15.4 mo) prior to surgery.
Procedure The study procedures were approved by the institutional review board of the hospital. Families were informed about the study by a research assistant during a preoperative visit, and written child assent, and parental permission was obtained from those interested in participating. Children and their primary caregivers then independently completed baseline study measures during a scheduled study visit occurring approximately 2 weeks prior to surgery (range, 1–4 wk). After the child’s surgery, a research assistant administered the perioperative (inpatient) self-report measures twice daily (between 7 AM and 10 AM and between 4 PM and 7 PM) until the child’s hospital discharge. Children received standard perioperative and postoperative pain management per institution protocol. Specifically, children received intrathecal morphine (15 μg/kg) intraoperatively. Postoperatively, children received continuous infusion of hydromorphone (4 µg/kg/hr), or morphine 0.2 mg/kg if hydromorphone was not tolerated, with a patient-controlled analgesia dose of 4 µg/kg and 6-minute lockout and bolus dose of 10 µg/kg for pain ratings above 5/10 up to once every 2 hours. Intravenous ketorolac boluses (0.5 mg/kg) up to once every 6 hours also were used for Spine
Pain Intensity Data on typical and highest back pain intensity were collected by visual analogue scales prior to surgery (as baseline predictors of postoperative pain outcomes), twice daily throughout the postoperative hospital stay until discharge (as perioperative predictors of subsequent pain outcomes), and at each of the 4 follow-up assessments. Children were instructed to mark anywhere along a 100-mm horizontal line with anchors “no pain” to “most pain” to indicate their typical and highest back pain intensity during the past month (at baseline and follow-up) or during the past 4 hours (during the postoperative hospital stay). Intensity scores were derived by measuring the distance in millimeters between the left anchor and where the child’s mark crossed the horizontal line; possible scores were in the range from 0 to 100, with higher scores indicating greater pain intensity. The visual analogue scale is the recommended measure of pain intensity for research with adolescents and is considered to be empirically well established.8,9 Anxiety Level of preoperative anxiety was measured at the baseline visit using the state portion of the State-Trait Anxiety Inventory for Children. The questionnaire contains 20 items for which children are asked to select adjectives that vary in intensity (e.g., “very calm,” “calm,” “not calm”) to best describe how they feel “right now.” The State-Trait Anxiety Inventory for Children is frequently used as an index of preoperative anxiety in children and has extensive support for its validity and internal consistency.10,11 Item scores were reverse coded where applicable and summed to comprise a total score, with higher scores indicating greater baseline anxiety. Negative Mood Mood was measured at all study time points (at baseline, twice daily throughout the hospital stay until discharge, and at each of the 4 follow-up assessments) using a 7-item semantic differential scale.12 Participants were presented with 7 items that varied in mood intensity along a 7-point scale (e.g., “extremely happy” to “extremely unhappy”) and were asked to select a response that best represented how they felt. Item www.spinejournal.com
E175
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE130580.indd E175
30/12/13 2:53 PM
DEFORMITY scores were summed to form a total score, with higher scores indicating greater negative mood. Pain Coping Efficacy Children’s confidence in ability to control pain was assessed using the following question: “Based on all the things you have done to deal with your pain, how confident are you that you could decrease your pain if you tried?”13 Children recorded their response on a 100-mm visual analogue scale with anchors “no confidence” to “complete confidence;” possible scores were in the range from 0 to 100 with higher scores indicating greater confidence in ability to control pain. Demographic Information Information about the age, sex, and race of the child and income level of the family (as a proxy for socioeconomic status) was obtained at baseline by a parent-report questionnaire. Medical Information Information on time since initial diagnosis of AIS, severity of spinal curvature (Cobb angle) at baseline, body mass index, length of fusion (in levels), lowest instrumented vertebrae, and duration of spinal fusion surgery was obtained through medical record review. Analgesic Use Data on daily (24-hour) cumulative opioid analgesic use during the postoperative hospital stay were documented by a research nurse and then averaged across hospital days for use in analyses. The medication data were converted to parenteral morphine-equivalent milligram dosages divided by weight using standard conversion ratios14 to account for multiple drug types and modes of delivery. Whether or not opioids continued to be used at the time of follow-up visits also was recorded.
Analyses Data on study measures were summarized for descriptive purposes using means, variances, and/or frequency distributions. Multilevel growth models with maximum likelihood estimation15 were then specified using HLM software (Scientific Software International Inc., Skokie, IL)16 to test the primary research questions. These models estimate changes over time in a given outcome variable (in this case pain intensity) for each individual and then estimate the overall average change (the intercept coefficient) and the extent to which specified predictor variables shift the average change trajectory (slope coefficients). An unconditional growth model (containing only “time” as a predictor variable) was used to derive an estimate of the typical daily rate of change in typical and highest pain intensity since surgery. Separate conditional growth models were then used to evaluate individual differences in the daily rate of change in pain intensity during the study time period as a function of demographic variables (age, sex, income level), baseline predictor variables (pain level, E176
www.spinejournal.com
Predictors of Pain in AIS • Connelly et al
Cobb angle, body mass index, time since diagnosis, anxiety, negative mood, and pain coping efficacy), and perioperative predictor variables (length of surgery, length of fusion, lowest instrumented vertebrae, postoperative pain levels, analgesic use, negative mood, and pain coping efficacy). Opioid analgesic use at the time of each study visit was entered into models as a covariate to control for the expected association between continued postoperative pain and continued use of analgesics. In addition, the multilevel growth models were specified to evaluate associations between the baseline predictor variables and preoperative pain; this enabled a determination of whether predictors of postoperative pain trajectories were different from variables already associated with pain at baseline.
RESULTS Descriptives Table 1 presents descriptive statistics for the mean pain intensity scores by time point. Almost all children (89%) reported at least mild pain intensity prior to surgery. Typical and highest pain intensity were on average mild to moderate at baseline (mean = 36/100 ± 25.51, and 50/100, respectively), moderate to severe during the hospital stay (mean = 49/100 and 64/100, respectively) and through 2-week follow-up (mean = 46/100 and 65/100, respectively), and then dropped to mild for the remainder of the follow-up periods (mean = 12–20/100 and mean = 23–38/100, respectively). Approximately 1 in 5 children (22%) reported greater pain intensity levels at their final follow-up assessment relative to their baseline (preoperative) levels. Two patients (4%) were on analgesics preoperatively. Most patients (70%) discontinued use of opioid analgesics by the 6-week postoperative follow-up.
Typical Trajectory of Postoperative Pain Figure 1 shows the mean typical and highest pain intensity from baseline through the follow-up period. Both typical and highest pain intensity on average gradually and significantly improved over days since surgery (b = −0.14 ± 0.02, t (49) = 12.33, P < 0.01, and b = −0.19 ± 0.03, t (49) = 6.78, P < 0.01, respectively).
Baseline Predictors of Preoperative Pain None of the demographic, medical, or psychological variables were found through multilevel model analyses to be significantly associated with preoperative typical pain intensity. However, baseline pain coping efficacy was inversely associated with highest pain intensity reported at baseline, b = −0.22 ± 0.10, t (46) = 2.32, P = 0.02.
Baseline Predictors of Postoperative Pain Trajectories Table 2 presents the results of the hierarchical linear models comprising baseline predictors of postoperative pain trajectories. Pain score trajectories did not reliably differ as a function of baseline Cobb angle, body mass index, time since diagnosis, or demographic variables. Children reporting higher baseline pain had a significantly slower daily rate February 2014
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE130580.indd E176
30/12/13 2:53 PM
DEFORMITY
Predictors of Pain in AIS • Connelly et al
TABLE 1. Pain Intensity Descriptive Statistics by Study Time Point Time Point Baseline
Hospital stay*
2-week follow-up
6-week follow-up
3-month follow-up
6-month follow-up
Measure
Mean ± SD
Typical pain level
36.10 ± 24.80
0–98
Highest pain level
49.58 ± 30.39
0–99
Typical pain level
49.47 ± 14.74
18–81
Highest pain level
63.99 ± 15.57
28–90
Typical pain level
45.56 ± 23.23
3–86
Highest pain level
65.29 ± 27.07
8–100
Typical pain level
20.12 ± 20.62
0–78
Highest pain level
37.76 ± 29.59
0–100
Typical pain level
13.91 ± 18.85
0–70
Highest pain level
24.53 ± 28.21
0–83
Typical pain level
11.57 ± 17.56
0–82
Highest pain level
23.48 ± 29.28
0–100
Range
*Obtained on the basis of twice daily assessment averaged over the length of the hospital stay after spinal fusion surgery. SD denotes standard deviation.
of improvement in typical and highest pain intensity, b = 0.002 ± 0.001, t (123) = 3.68, P < 0.01, and b = 0.001 ± 0.001, t (123) = 2.14, P = 0.03, respectively. Children with greater pain coping efficacy at baseline had a significantly more rapid improvement in postoperative typical and highest pain intensity, b = −0.002 ± 0.001, t (123) = −2.64, P < 0.01, and b = −0.002 ± 0.001, t (123) = −2.13, P = 0.04. In addition, children with higher baseline anxiety, but not children with higher baseline negative mood, had a significantly slower daily rate of improvement in postoperative typical and highest pain, b = 0.01 ± 0.003, t (122) = 2.37, P = 0.02, and b = 0.001 ± 0.004, t (122) = 2.37, P = 0.05. Figure 2 and Figure 3 display postoperative pain trajectories as a function of highest and lowest quartile baseline anxiety and pain scores, respectively.
Figure 1. Pain intensity ratings by study time point. FU indicates follow-up. Spine
Perioperative Predictors of Postoperative Pain Trajectories Table 3 presents the results of the hierarchical linear growth models comprising the perioperative predictors of postoperative pain trajectories. Only pain coping efficacy reported during the hospital stay predicted more rapid improvement in typical and highest pain trajectories, b = −0.002 ± 0.001, t (123) = −2.18, P = 0.03, and b = −0.002 ± 0.001, t (123) = −2.29, P = 0.02. Length of surgery, length of fusion, lowest instrumented vertebrae, amount of opioid analgesic use during the hospital stay, and mood ratings during the hospital stay all were unrelated to postoperative pain trajectories.
DISCUSSION This study sought to determine the extent to which certain demographic, biomedical, and psychological variables measured at baseline and/or immediately after spinal fusion surgery for AIS predict differences in the course of adolescents’ pain during the initial 6-month period after the procedure. Pain on average gradually and significantly declined throughout the time interval studied. However, the rate of decline in pain levels was significantly slower for adolescents who prior to surgery reported high levels of pain and anxiety, and approximately 1 in 5 adolescents reported greater pain intensity levels throughout their follow-up assessment compared with their preoperative pain level. Children who prior to surgery and throughout the hospital stay had greater confidence in their ability to cope with pain had lower pain intensity preoperatively and had more rapid improvement in pain levels through 6 months after surgery. Demographic and biomedical variables did not reliably predict differences in pain trajectories. www.spinejournal.com
E177
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE130580.indd E177
30/12/13 2:53 PM
DEFORMITY
Predictors of Pain in AIS • Connelly et al
TABLE 2. Results of Linear Growth Modeling Predicting Daily Changes in Pain Since Surgery From
Baseline Predictors
For Changes in Typical Pain Predictor
b (±SE)
Intercept*
−0.14 (±0.02)
Sex
T
For Changes in Highest Pain P
b (±SE)
P
−7.60
