- Email: [email protected]
Long-Term Effects of Neonatal Morphine Infusion on Pain Sensitivity: Follow-Up of a Randomized Controlled Trial
The Journal of Pain, Vol 16, No 9 (September), 2015: pp 926-933 Available online at www.jpain.org and www.sciencedirect.com
Long-Term Effects of Neonatal Morphine Infusion on Pain Sensitivity: Follow-Up of a Randomized Controlled Trial Abraham J. Valkenburg,* Gerbrich E. van den Bosch,* Joke de Graaf,* Richard A. van Lingen,y Nynke Weisglas-Kuperus,z Joost van Rosmalen,x Liesbeth J. M. Groot Jebbink,y Dick Tibboel,* and Monique van Dijk*,z Departments of *Pediatric Surgery and zNeonatology, Erasmus University Medical Center–Sophia Children’s Hospital, Rotterdam, The Netherlands. y Princess Amalia Department of Pediatrics, Division of Neonatology, Isala Clinics, Zwolle, The Netherlands. x Department of Biostatistics, Erasmus University Medical Center, Rotterdam, The Netherlands.
Abstract: Short-term and long-term effects of neonatal pain and its analgesic treatment have been topics of translational research over the years. This study aimed to identify the long-term effects of continuous morphine infusion in the neonatal period on thermal pain sensitivity, the incidence of chronic pain, and neurological functioning. Eighty-nine of the 150 participants of a neonatal randomized controlled trial on continuous morphine infusion versus placebo during mechanical ventilation underwent quantitative sensory testing and neurological examination at the age of 8 or 9 years. Forty-three children from the morphine group and 46 children from the placebo group participated in this follow-up study. Thermal detection and pain thresholds were compared with data from 28 healthy controls. Multivariate analyses revealed no statistically significant differences in thermal detection thresholds and pain thresholds between the morphine and placebo groups. The incidence of chronic pain was comparable between both groups. The neurological examination was normal in 29 (76%) of the children in the morphine group and 25 (61%) of the children in the control group (P = .14). We found that neonatal continuous morphine infusion (10 mg/kg/h) has no adverse effects on thermal detection and pain thresholds, the incidence of chronic pain, or overall neurological functioning 8 to 9 years later. Perspective: This unique long-term follow-up study shows that neonatal continuous morphine infusion (10 mg/kg/h) has no long-term adverse effects on thermal detection and pain thresholds or overall neurological functioning. These findings will help clinicians to find the most adequate and safe analgesic dosing regimens for neonates and infants. ª 2015 by the American Pain Society Key words: Morphine, quantitative sensory testing, neonatal intensive care, follow-up, randomized controlled trial.
P
roviding adequate and evidence-based analgesia and sedation to neonates receiving intensive care is an ongoing challenge; one has to account for developmental changes in drug pharmacokinetics and
Received March 25, 2015; Revised May 29, 2015; Accepted June 15, 2015. No funding sources were provided. None of the authors report financial relationships that are relevant to this work. None of the authors report a conflict of interest. Supplementary data accompanying this article are available online at www.jpain.org and www.sciencedirect.com. Address reprint requests to Abraham J. Valkenburg, MD, PhD, Intensive Care and Department of Pediatric Surgery, Erasmus University Medical Center–Sophia Children’s Hospital, Dr. Molewaterplein 60, 3015 GJ Rotterdam, The Netherlands. E-mail: [email protected] 1526-5900/$36.00 ª 2015 by the American Pain Society http://dx.doi.org/10.1016/j.jpain.2015.06.007
926
pharmacodynamics as well as in the human nervous system.13,23 The repeated painful procedures in intensive care can lead to short-term hyperalgesia.22 Long-term follow-up of extremely preterm neonates showed a generalized decrease in thermal sensitivity, probably caused by tissue injury and modulation of nociceptor pathways.26 Therefore, analgesic treatment plays an important role in the neonatal intensive care unit (NICU). Opioids such as morphine are commonly used in the NICU and have both beneficial and adverse short-term effects. Morphine is an effective analgesic agent for neonates’ postoperative pain but not for acute procedural pain.1,4,5,20 However, continuous morphine infusion does not decrease the risk of poor neurological outcome after intensive care treatment.20
Valkenburg et al Animal studies on long-term cognitive functioning and neurodevelopmental outcome after neonatal morphine exposure found impaired learning abilities, neuronal degeneration,2 increased neuroapoptosis,3 and enhanced hippocampal gliosis (proliferation of damaged astrocytes) in rodents treated with morphine.14,21,25 In humans, participants of 2 randomized controlled trials (RCTs) on neonatal morphine administration were studied 5 years after the original RCT was conducted.1,20 The first, a small-scale follow-up of the NEOPAIN (Neurologic Outcomes and Pre-emptive Analgesia in Neonates) trial,9 showed that children who had received morphine (n = 14) had a smaller head circumference, weighed less, and had more social problems than children who had received placebo (n = 5). The second, performed in our institution, showed that children who had received morphine (n = 49) performed more poorly on 1 subtest of the intelligence scale than did the children who had received placebo (n = 41); other neurobehavioral outcomes and the incidences of chronic pain were comparable between the 2 groups.7 This unique cohort is being followed, and at the age of 8 years, participants were old enough for quantitative sensory testing (QST).6 Morphine is used worldwide for opioid analgesia in neonates, infants, and children; therefore, it is important to confirm that it does not have any negative long-term effects. In the current study, we evaluated thermal sensitivity, the occurrence of chronic pain, and neurological outcome in 8- to 9-year-old children who at neonatal age had received continuous morphine infusion. Our null hypothesis is that continuous morphine infusion at neonatal age has no adverse effects on thermal detection and pain thresholds, incidence of chronic pain, and neurological functioning at 8 to 9 years of age.
Methods Original Study Between 2000 and 2002, 150 term- and preterm-born neonates who received mechanical ventilation in 2 level III NICUs participated in a multicenter RCT. Morphine is one of the most frequently used analgesics across all age groups worldwide. At the time of the study,20 the debate was whether preterm newborns on ventilatory support should routinely receive a continuous morphine infusion. The aim of the RCT was to evaluate the effects of continuous intravenous morphine infusion on pain responses, incidence of intraventricular hemorrhage, and poor neurologic outcome (severe intraventricular hemorrhage, periventricular leukomalacia, or death). Seventy-three neonates were randomly assigned to the continuous morphine group (loading dose of 100 mg/kg followed by infusion of 10 mg/kg/h) and 77 to the placebo group (normal saline infusion). If pain or distress was noted, children in both groups received an open-label morphine bolus of 50 mg/kg and, if indicated, an openlabel morphine infusion (5–10 mg/kg/h) as rescue medication. Open-label morphine was administered to 27% of the children in the morphine group versus 40% of the
The Journal of Pain
927
placebo group (P = .10). Further details, including the background characteristics of the participants, can be found in the original article.20
Follow-Up Study (8–9 Years) The institutional ethics review boards of the 2 study sites (Erasmus University Medical Center–Sophia Children’s Hospital, Rotterdam, The Netherlands, and Isala Clinics, Zwolle, The Netherlands) approved the study plan. Parents of the 132 survivors were informed of the study and were asked for written informed consent. Seventeen participants from the morphine group were lost to follow-up and 5 parents refused to provide informed consent for the follow-up study. Sixteen patients of the placebo group were lost to follow-up and 5 parents refused informed consent for the follow-up study. The remaining 89 children and their parents were then invited for a follow-up visit in their hospital (either Rotterdam or Zwolle) (Fig 1). The nonparticipants had a lower birth weight and their first hospital stay was longer compared with the participants.8 Parents were asked to complete several questionnaires (see later discussion) before the visit. The visit consisted of 3 parts: QST by a trained researcher (A.J.V.), medical examination by a pediatrician (D.T.), and neuropsychological testing by a psychologist (M.vD.). These health professionals were blinded to the participants’ study condition (continuous morphine infusion versus placebo) in the original RCT.
QST Participants underwent QST (see Supplementary Material) in a quiet hospital room, with a stable room temperature (20–22 C). Parents were present in the room and were instructed not to interfere during the test. Reaction time was measured using the baseline speed task for the dominant hand (Amsterdam Neuropsychological Tasks, Version 3.1; Boom Test Publishers, Amsterdam, The Netherlands).
Reference Data QST (Control Group) We compared the detection and pain thresholds of both the placebo and the morphine group with detection and pain thresholds of 28 healthy control individuals (aged 8–10 years). The control values were collected by our research group and obtained using the same protocol as for QST.24 Healthy term-born controls without a history of intensive care admission or severe early pain were recruited. Participants of other studies were asked whether they could recommend someone in the age range of 8– 18 years who would be interested in volunteering. Furthermore, children were recruited at primary schools in Rotterdam. Eventually, we selected the children aged 8 to 10 years as the reference group for the present study. Control individuals were included at Erasmus Medical Center in Rotterdam from October 2011 to March 2013. The local institutional review board approved this study. Informed consent from the parents and assent of the children was obtained before participation.
928
Long-Term Effects of Neonatal Morphine Infusion
The Journal of Pain
Figure 1. Flowchart based on the CONSORT checklist.
Chronic Pain Questionnaire
Statistical Analysis
Parents of all children also completed the Chronic Pain Questionnaire (Dutch version)18; this questionnaire examines the incidence and characteristics of pain in the 3 months before the visit (location, frequency, duration, and intensity). Chronic pain is defined as a pain episode with duration of 3 months or longer (either recurrent pain with pain-free intervals or continuous pain). The questionnaire was developed and validated for children between 4 and 18 years old and can be completed in less than 10 minutes.
Summary statistics are presented as medians (interquartile range [IQR]) for continuous variables and as percentages for categorical variables. Data were compared between the continuous morphine group and the placebo group using the Mann-Whitney test for continuous, nonnormal data and the c2 test (or Fisher exact test in the case of low predicted cell counts) for nominal data. IQ data (Wechsler Intelligence Scale for Children) are presented as means (standard deviation [SD]) and compared between the morphine and placebo groups using the independent samples t-test. The QST data of the morphine and placebo groups were also compared with data of 28 healthy control individuals using analysis of covariance, correcting for age using a post hoc test with Bonferroni correction for multiple comparisons. For the multivariate analysis of the QST data, we used robust regression analysis with MM estimation28 to account for the fact that the model residuals were not normally distributed. The weight function was the Tukey bisquare estimator. A model was built for each of the 6 QST modalities (see Supplementary Material), with treatment condition (morphine vs placebo), total amount of morphine in the first 28 days of life, age, sex, study site, and propensity score as predictor variables. The propensity score was derived from logistic regression analysis in which the treatment condition (morphine or placebo) was the criterion and the covariables were study site, sex, gestational age, birth weight, Clinical Risk Index for Babies, and socioeconomic status. In this way, several
Medical Examination All children were examined by a pediatrician (D.T.) who referred them for further diagnosis or treatment if indicated. Weight, height, and head circumference were measured and plotted against sex-matched reference values for The Netherlands (fourth nationwide growth study, 1997).10 The neurological examination was based on the Touwen assessment of minor neurological dysfunctions.15 We assessed 41 items in 5 domains (posture/muscle tone, reflexes, involuntary movements, coordination/balance, and cranial nerve dysfunctions). Minor neurological dysfunctions were defined as the presence of 2 or more deviant items in at least 1 domain.
Neuropsychological Testing A trained psychologist (M.vD.) tested the participants’ IQ with the Wechsler Intelligence Scale for Children-III (Dutch version).16
Valkenburg et al
The Journal of Pain
relevant covariables are combined into 1 weighted score.7,8,27 Logistic regression analyses were applied with prevalence of chronic pain and minor neurological dysfunctions (0 or 1 vs >1) as dichotomous outcome variables and with the treatment condition (continuous morphine vs placebo) and the total amount of morphine in the first 28 days after birth as predictor variables. P values (2-sided) of less than .05 were considered statistically significant. Data were analyzed with SPSS software, version 19.1 (IBM Corp, Armonk, NY) and SAS, version 9.3 (SAS Institute Inc, Cary, NC).
Results Background Characteristics Of the 132 survivors, 43 children from the morphine group and 46 children from the placebo group participated in this follow-up study (Fig 1). Background characteristics were comparable between the 2 groups (Table 1). Total morphine exposure in the first 28 days of life was statistically significantly higher in the morphine group (P < .001) than in the placebo group. Twelve children (26.1%) in the placebo group did not receive any opioids. Mean IQ was 101 (SD = 18) for the placebo group versus 99 (SD = 19) for the morphine group (P = .63). Twenty-eight controls were included to compare thermal and pain sensitivity as well as incidence of chronic pain among the morphine group, placebo group, and healthy controls.
QST QST was performed in 41 children (89%) of the placebo group, 37 children (86%) of the morphine group, and 28 healthy control individuals. Nine children did not attend
Table 1.
929
the follow-up visit (however, parents did complete the questionnaires), and the equipment was not available on the remaining 2 occasions. QST was feasible in all participants with regard to understanding the instructions and completing the test. No significant differences in detection or pain thresholds were found among the placebo group, morphine group, and the control group using both the method of limits and the method of levels (Table 2). Ten (24%) of the children in the placebo group did not establish a pain threshold for cold before the minimum temperature (0 C) was reached versus 9 (25%) of the morphine group and 16 (57%) of the control individuals (P = .006; post hoc test: significant difference between both placebo group and morphine group versus controls). Ten (24%) of the children in the placebo group did not establish a pain threshold for heat before the maximum temperature (50.0 C) was reached, versus 8 (22%) of the children in the morphine group and 13 (46.4%) of the control individuals (P = .38). Regression estimates for the 6 QST modalities are presented in Tables 3–5. Neither treatment condition (morphine group vs placebo) nor the total amount of morphine in the first 28 days was a statistically significant predictor in any of the models. For the cold and warmth detection thresholds (method of limits), IQ was a statistically significant covariate (P = .03 and P = .01, respectively). The higher the IQ, the more sensitive the children were for the detection of cold and warmth (Table 3). Children tested in the hospital in Zwolle (n = 39) had statistically significant lower mean cold pain thresholds (P < .001) and borderline statistically significant lower mean heat pain thresholds (P = .054) (Table 4). For the warm detection thresholds (method of levels), IQ was a statistically significant covariate (P = .007).
Background Characteristics (Distinguished by Group)
Characteristics Male sex Birth characteristics Gestational age, wk Birth weight, g Age, y Height, cm SDS height for agey Weight, kg SDS weight for agey Head circumference, cm SDS head circumference for agey IQz Reaction time, msx Test location Rotterdam Zwolle
PLACEBO GROUP (N = 46) 28 (61%) 31 (28–32) 1520 (995–2145) 8.9 (8.7–9.3) 137 (131–140) .02 (.7 to 1.1) 29.2 (26.5–32.8) .05 (.6 to .7) 52 (51–53) .20 (.8 to .7) 101 (18) 391 (354–442) 25 (54%) 21 (46%)
Abbreviation: SDS, standard deviation score. NOTE. Values are median (IQR) or number (%). *Comparison Mann-Whitney test for continuous variables and Fisher exact test for categorical variables. yAccording to reference values (The Netherlands, 1997). zIQ (Wechsler Intelligence Scale for Children), mean (SD), P value from t-test. xAmsterdam Neuropsychological Tasks.
MORPHINE GROUP (N = 43) 28 (65%) 30 (29–32) 1400 (1024–1788) 8.9 (8.8–9.1) 135 (131–140) .04 (.5 to .7) 27.9 (24.6–31.6) .01 (1.3 to .8) 52 (51–54) .03 (1.0 to .6) 99 (19) 390 (331–442) 25 (58%) 18 (42%)
P VALUE* .83 .39 .70 .54 .42 .72 .29 .53 .90 .92 .63 .82 .83
930
Long-Term Effects of Neonatal Morphine Infusion
The Journal of Pain
Table 2.
Results of QST, Distinguished by Group Modality
Method of limits Cold detection threshold, C Warm detection threshold, C Cold pain threshold, C Threshold not reached at 0 Cy Heat pain threshold, C Threshold not reached at 50 Cy Method of levels Cold detection threshold, C Number of stimuli Warm detection threshold, C Number of stimuli
PLACEBO GROUP (N = 41)
MORPHINE GROUP (N = 37)
CONTROL GROUP (N = 28)
P VALUE*
29.6 (2.4) 34.8 (1.4) 11.8 (8.4) 10 (24%) 45.0 (3.5) 10 (24%)
29.9 (1.4) 34.5 (1.4) 12.3 (8.9) 9 (25%) 44.9 (4.3) 8 (22%)
30.5 (1.0) 34.2 (1.5) 9.3 (10.1) 16 (57%) 45.2 (5.0) 13 (46%)
.15 .23 .40 .006z .74 .38
30.7 (1.2) 11 (3) 33.3 (1.1) 11 (4)
31.1 (.7) 10 (2) 33.2 (1.1) 10 (2)
30.5 (1.5) 11 (3) 33.6 (.9) 10 (3)
.27 .38 .84 .30
NOTE. Values are mean (SD) or number (%). *Analysis of covariance test for the comparison among the 3 groups corrected for age. yPatients in whom 1 or more times the pain threshold was not reached (the child did not press the button before the temperature of the thermode reached its minimum/ maximum (.0 C and 50.0 C, respectively). zPost hoc Bonferroni correction: placebo group and morphine group versus control individuals, P < .05.
Chronic Pain Chronic pain (duration longer than 3 months) was reported in 4 (9%) children in the placebo group versus 5 (12%) in the morphine group (P = .42). Abdominal pain was the most common type of pain; it was present in 3 of the children in the placebo group and 3 children in the morphine group (P = 1.00). The second most common type of pain was headache; it was present in 4 of the children in the placebo group versus 2 in the
Robust Regression Estimates of Intervention, With Adjustment for Different Covariates: Detection Thresholds (Method of Limits)
Table 3.
Detection Threshold
ESTIMATE
Detection threshold for cold Intercept 31.71 Treatment condition* .14 Total morphiney .08 Sex .18 Age .31 Propensity scorez .05 Study site .03 IQx .01 Detection threshold for warmth Intercept 32.28 Treatment condition* .35 Total morphiney .12 Sex .38 Age .45 Propensity scorez .32 Study site .47 IQx .02
95% CI LIMITS
P VALUE
26.04 to 37.39 .32 to .60 .12 to .29 .62 to .26 .93 to .31 .53 to .63 .40 to .46 .001 to .02
<.001 .55 .43 .43 .32 .86 .88 .03
23.75 to 40.80 1.01 to .31 .43 to .20 .27 to 1.03 .48 to 1.39 1.15 to .50 .15 to 1.11 .039 to .004
<.001 .30 .47 .25 .34 .44 .14 .01
Abbreviation: CI, confidence interval. *Morphine group versus placebo group. yTotal morphine in first 28 d. zPropensity score based on neonatal characteristics (study location, sex, duration mechanical ventilation, gestational age, Clinical Risk Index for Babies, and socioeconomic status [middle and high vs low]). xIQ (Wechsler Intelligence Scale for Children).
morphine group (P = .17). The intensity of the chronic pain was low; the median (IQR) visual analog scale score was 1.00 (.50–2.05) for the morphine group versus 1.85 (.63–4.20) in the placebo group (P = .41). Two children in the morphine group complained of pain 2 or 3 times a month, and 3 children complained 2 or more times a week. Two children in the placebo group complained of pain once a month, and 2 children complained 2 or more times a week (P = .91). In the control group, chronic pain was reported in 5 (18%) of the children.
Robust Regression Estimates of Intervention, With Adjustment for Different Covariates: Pain Thresholds (Method of Limits)
Table 4.
Pain Threshold Pain threshold for cold Intercept Treatment condition* Total morphiney Sex Age Propensity scorez Study site IQx Pain threshold for warmth Intercept Treatment condition* Total morphiney Sex Age Propensity scorez Study site IQx
ESTIMATE
95% CI LIMITS
P VALUE
.95 1.91 .25 .51 1.44 .24 7.78 .02
57.77 to 59.66 2.80 to 6.61 2.05 to 2.54 5.11 to 4.09 4.97 to 7.85 5.89 to 5.41 12.22 to 3.35 .10 to .14
.97 .43 .83 .83 .66 .93 <.001 .77
46.71 .32 .05 1.16 .22 .52 1.94 .04
20.47 to 72.94 1.79 to 2.42 1.08 to .99 3.20 to .88 2.65 to 3.10 3.06 to 2.02 .04 to 3.91 .09 to .01
<.001 .77 .93 .26 .88 .69 .05 .15
Abbreviation: CI, confidence interval. *Morphine group versus placebo group. yTotal morphine in first 28 d. zPropensity score based on neonatal characteristics (study location, sex, duration mechanical ventilation, gestational age, Clinical Risk Index for Babies, and socioeconomic status [middle and high vs low]). xIQ (Wechsler Intelligence Scale for Children).
Valkenburg et al
The Journal of Pain
Robust Regression Estimates of Intervention, With Adjustment for Different Covariates: Detection Thresholds (Method of Levels)
Table 5.
Detection Threshold
ESTIMATE
Detection threshold for cold Intercept 34.66 Treatment condition* .31 Total morphiney .03 Sex .05 Age .48 Propensity scorez .27 Study site .13 IQx .004 Detection threshold for warmth Intercept 31.01 Treatment condition* .18 Total morphiney .07 Sex .17 Age .38 Propensity scorez .42 Study site .20 IQx .01
95% CI LIMITS
P VALUE
28.74 to 40.59 .13 to .75 .22 to .17 .46 to .36 1.12 to .16 .79 to .25 .28 to .54 .007 to .02
<.001 .16 .80 .81 .14 .31 .53 .46
26.45 to 35.57 .54 to .19 .10 to .23 .18 to .51 .11 to .87 .04 to .89 .53 to .14 .02 to .003
<.001 .35 .44 .35 .13 .07 .25 .007
Abbreviation: CI, confidence interval. *Morphine group versus placebo group. yTotal morphine in first 28 d. zPropensity score based on neonatal characteristics (study location, sex, duration mechanical ventilation, gestational age, Clinical Risk Index for Babies, and socioeconomic status [middle and high vs low]). xIQ (Wechsler Intelligence Scale for Children).
Logistic regression analysis revealed that neither treatment condition (morphine vs placebo) nor the total amount of morphine in the first 28 days was a significant predictor for the presence of chronic pain (P = .68 vs P = .84).
Neurological Examination Two children in the placebo group and 2 children in the morphine group were severely intellectually and developmentally disabled and did therefore not attend the follow-up visit. Neurological examination was performed in 38 children of the morphine group and 41 of the placebo group (Table 6). The neurological examination was normal in 29 (76%) of the children in the morphine group and 25 (61%) of the children in the control group (P = .14). Logistic regression revealed that neither group (morphine vs placebo) nor the total Table 6.
931
amount of morphine in the first 28 days was a significant predictor for the presence of minor neurological dysfunctions (P = .30 vs P = .23). The presence of minor neurological dysfunctions at 8 to 9 years of age was not related to the presence of intraventricular hemorrhages in the neonatal period (P = .26).
Discussion A cohort of children who as neonates participated in an RCT on continuous morphine infusion versus placebo for mechanical ventilation was studied at the age of 8 to 9 years. In the multivariate analyses, we found that the treatment condition (morphine vs placebo) was not a significant predictor in any of the 6 QST modalities. The multivariate analysis did show correlations between IQ and the detection thresholds, with both the reaction time–dependent method of limits and the reaction time–independent method of levels. However, all correlations were weak and clinically not significant. In another study by our group, we found that in 5-yearold children, detection threshold levels assessed with the method of limits were significantly correlated with IQ (r = .64; P = .006), whereas the detection threshold levels assessed with the method of levels were not correlated with IQ.6 However, in the present study, we also found a weak correlation between IQ and the thermal detection thresholds assessed with the reaction time–independent method of levels. This suggests that children with a higher IQ could have a better attentiveness as well.
Long-Term Effects of Surgery and Morphine Hermann et al12 reported long-term hypoalgesia for heat pain in both preterm and term-born 9- to 14-year-old children who had received neonatal intensive care. The preterm neonates in that study underwent a mean number of 172 invasive procedures in the first week of life; however, no more than 53% of them received analgesics. The hypoalgesia in former extremely preterm-born 11-year-olds, reported by Walker et al,26 was most marked in those who had undergone neonatal surgery. In contrast, another study on long-term effects of neonatal surgery (all neonates received opioids postoperatively) showed no differences in the cold and warmth
Minor Neurological Dysfunctions (Distinguished by Group) Dysfunction
PLACEBO GROUP (N = 41)
MORPHINE GROUP (N = 38)
P VALUE*
Minor neurological dysfunctions Mild deviations in posture/muscle tone Mild deviations in reflexes Presence of involuntary movements Mild deviations in coordination/balance Mild deviations in cranial nerve function
16 (39) 6 (15) 4 (10) 0 (0) 5 (12) 1 (2)
9 (24) 2 (5) 1 (3) 0 (0) 7 (18) 0 (0)
.14 .27 .36 1.00 .53 1.00
NOTE. Values are number (%). *Fisher exact test.
932
The Journal of Pain
detection thresholds between the neonatal surgery group and the control group at the age of 9 to 12 years.19 Combining the previous data with the results of the present study, we hypothesize that neonatal injury or surgery is likely to have more pronounced long-term effects on pain processing than neonatal morphine treatment itself.
Chronic Pain The prevalence of chronic pain was comparable between the 2 groups and lower than the reference values (ie, 11% for boys vs 13% for girls).18 The prevalence of chronic pain in both groups had decreased since the 5-year follow-up visit, ie, from 14% to 9% in the placebo group and from 15% to 12% in the morphine group.7
Neurological Functioning The results of the neurological examination were normal for most of the children. The prevalence of minor neurological dysfunctions (ie, 39% of the placebo group vs 24% of the morphine group) is comparable with the prevalence in a reference cohort of term-born children (ie, 50% prevalence of minor neurological dysfunctions).15 Mild deviations in coordination/balance control (a minor neurological dysfunction that could indicate cerebellar dysfunction) were more common in the morphine group in the present study. Research in rodents suggests that morphine treatment has a negative effect on the development of cerebellar neurons.11 A study in human preterm-born infants17 found that they were more at risk for cerebellar injuries during the neonatal period than were term-born infants and that at adolescent age their cerebellum volume was smaller than that in term-born adolescents. The minor neurological dysfunctions observed may reflect the combined consequences of prematurity and critical illness along with potential effects of morphine, not the effects of morphine in isolation. The pilot follow-up study of the NEOPAIN trial9 found an overall lower prevalence of neurological soft signs (ie, 20% in the placebo group [n = 5] vs 14% in the morphine group [n = 14]); however, the researchers did not provide details on which neurological soft signs (comparable with minor neurological dysfunctions) were assessed. In that study, the children
References 1. Anand KJ, Hall RW, Desai N, Shephard B, Bergqvist LL, Young TE, Boyle EM, Carbajal R, Bhutani VK, Moore MB, Kronsberg SS, Barton BA: Effects of morphine analgesia in ventilated preterm neonates: primary outcomes from the NEOPAIN randomised trial. Lancet 363:1673-1682, 2004 2. Atici S, Cinel L, Cinel I, Doruk N, Aktekin M, Akca A, Camdeviren H, Oral U: Opioid neurotoxicity: comparison of morphine and tramadol in an experimental rat model. Int J Neurosci 114:1001-1011, 2004 3. Bajic D, Proudfit HK: Projections from the rat cuneiform nucleus to the A7, A6 (locus coeruleus), and A5 pontine noradrenergic cell groups. J Chem Neuroanat 50-51:11-20, 2013
Long-Term Effects of Neonatal Morphine Infusion in the morphine group had 7% smaller head circumference and 4% less body weight than children in the placebo group. Although several children in our study were born prematurely or were born small for gestational age, most children in both groups now had normal height, weight, and head circumference for their age. The preterm neonates in the NEOPAIN trial received higher doses of morphine, up to 30 mg/kg/h of morphine.1 The children in the morphine group in our study received 10 mg/kg/h of morphine at neonatal age; in the case of pain or distress, children in both groups received open-label morphine as rescue medication.
Limitations Given the range of variation in data on pain thresholds in other studies4 and the fact that the study site (Zwolle vs Rotterdam) was a significant covariate for the cold pain thresholds in the present study, we suggest that assessing pain thresholds is sensitive to variations in instructions and methodology. However, this does not influence the comparison among the morphine group, placebo group, and control group in the present study. Children in all the study groups received the same instructions.
Conclusions We found in the present study that continuous morphine infusion (10 mg/kg/h) during the neonatal period had no adverse effects on thermal detection and pain thresholds, the incidence of chronic pain, or overall neurological functioning 8 to 9 years later. Clinicians around the world encounter postoperative and procedural pain in neonates on a day-to-day basis. These are important findings in view of the concerns on neurotoxicity of opioids and anesthetic agents and will help clinicians to find the most adequate and safe analgesic dosing regimens for neonates and infants.
Supplementary Data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jpain.2015.06.007.
4. Bellu’ R, Dewaal K, Zanini R: Opioids for neonates receiving mechanical ventilation. a systematic review and meta-analysis. Arch Dis Child-Fetal 95:241-251, 2009 5. Carbajal R, Lenclen R, Jugie M, Paupe A, Barton BA, Anand KJS: Morphine does not provide adequate analgesia for acute procedural pain among preterm neonates. Pediatrics 115:1494-1500, 2005 6. de Graaf J, Valkenburg AJ, Tibboel D, van Dijk M: Thermal detection thresholds in 5-year-old preterm born children. Early Hum Dev 88:487-491, 2012 7. de Graaf J, van Lingen RA, Simons SHP, Anand KJS, Duivenvoorden HJ, Weisglas-Kuperus N, Roofthooft DWE, Jebbink LJMG, Veenstra RR, Tibboel D, van Dijk M: Longterm effects of routine morphine infusion in mechanically ventilated neonates on children’s functioning: five-year
Valkenburg et al
The Journal of Pain
933
follow-up of a randomized controlled trial. Pain 152: 1391-1397, 2011
Wouden JC: Pain in children and adolescents: a common experience. Pain 87:51-58, 2000
8. de Graaf J, van Lingen RA, Valkenburg AJ, WeisglasKuperus N, Groot Jebbink L, Wijnberg-Williams B, Anand KJ, Tibboel D, van Dijk M: Does neonatal morphine use affect neuropsychological outcomes at 8 to 9 years of age? Pain 154:449-458, 2013
19. Schmelzle-Lubiecki BM, Campbell KA, Howard RH, Franck L, Fitzgerald M: Long-term consequences of early infant injury and trauma upon somatosensory processing. Eur J Pain 11:799-809, 2007
9. Ferguson SA, Ward WL, Paule MG, Hall RW, Anand KJ: A pilot study of preemptive morphine analgesia in preterm neonates: effects on head circumference, social behavior, and response latencies in early childhood. Neurotoxicol Teratol 34:47-55, 2012 10. Fredriks AM, van Buuren S, Wit JM, VerlooveVanhorick SP: Body index measurements in 1996-7 compared with 1980. Arch Dis Child 82:107-112, 2000 11. Hauser KF, Gurwell JA, Turbek CS: Morphine inhibits Purkinje cell survival and dendritic differentiation in organotypic cultures of the mouse cerebellum. Exp Neurol 130:95-105, 1994 12. Hermann C, Hohmeister J, Demirakc¸a S, Zohsel K, Flor H: Long-term alteration of pain sensitivity in schoolaged children with early pain experiences. Pain 125: 278-285, 2006 13. Johnston CC, Fernandes AM, Campbell-Yeo M: Pain in neonates is different. Pain 152:S65-S73, 2011 14. Juul SE, Beyer RP, Bammler TK, Farin FM, Gleason CA: Effects of neonatal stress and morphine on murine hippocampal gene expression. Pediatr Res 69:285-292, 2010 15. Kikkert HK, De Jong C, Hadders-Algra M: Minor neurological dysfunction and IQ in 9-year-old children born at term. Dev Disabil Res Rev 53:e16-25, 2011 16. Kort W, Schittekatte M, Dekker PH, Verhaeghe P, Compaan EL, Bosmans M, Vermeir G: WISC-III NL Wechsler Intelligence Scale for Children. In: Derde Editie NL (ed): Handleiding en Verantwoording. Amsterdam, Harcourt Test Publishers/Nederlands Instituut voor Psychologen, 2005
20. Simons SH, van Dijk M, van Lingen RA, Roofthooft D, Duivenvoorden HJ, Jongeneel N, Bunkers C, Smink E, Anand KJS, van den Anker JN, Tibboel D: Routine morphine infusion in preterm newborns who received ventilatory support: a randomized controlled trial. JAMA 290:2419-2427, 2003 21. Sternberg WF, Scorr L, Smith LD, Ridgway CG, Stout M: Long-term effects of neonatal surgery on adulthood pain behavior. Pain 113:347-353, 2005 22. Taddio A, Shah V, Atenafu E, Katz J: Influence of repeated painful procedures and sucrose analgesia on the development of hyperalgesia in newborn infants. Pain 144:43-48, 2009 23. Tibboel D, Anand KJS, van den Anker JN: The pharmacological treatment of neonatal pain. Semin Fetal Neonat Med 10:195-205, 2005 24. van den Bosch GE, White T, Tibboel D, van Dijk M: Functional MRI pain studies in children? Yes, we (s)can!. Pediatr Radiol 43:1235-1236, 2013 25. Vien TN, Gleason CA, Hays SL, McPherson RJ, Chavkin C, Juul SE: Effects of neonatal stress and morphine on kappa opioid receptor signaling. Neonatology 96:235-243, 2009 26. Walker SM, Franck LS, Fitzgerald M, Myles J, Stocks J, Marlow N: Long-term impact of neonatal intensive care and surgery on somatosensory perception in children born extremely preterm. Pain 141:79-87, 2009
17. Manto MU, Jissendi P: Cerebellum: links between development, developmental disorders and motor learning. Front Neuroanat 6:1-10, 2012
27. Wright AA, Zhang B, Ray A, Mack JW, Trice E, Balboni T, Mitchell SL, Jackson VA, Block SD, Maciejewski PK, Prigerson HG: Associations between end-of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment. JAMA 300:1665-1673, 2008
18. Perquin CW, Hazebroek-Kampschreur AA, Hunfeld JA, Bohnen AM, van Suijlekom-Smit LW, Passchier J, van der
28. Yohai VJ, Zamar RH: Optimal locally robust M-estimates of regression. J Stat Plan Inference 64:309-323, 1997
Long-Term Effects of Neonatal Morphine Infusion on Pain Sensitivity: Follow-Up of a Randomized Controlled Trial Abraham J. Valkenburg,* Gerbrich E. van den Bosch,* Joke de Graaf,* Richard A. van Lingen,y Nynke Weisglas-Kuperus,z Joost van Rosmalen,x Liesbeth J. M. Groot Jebbink,y Dick Tibboel,* and Monique van Dijk*,z Departments of *Pediatric Surgery and zNeonatology, Erasmus University Medical Center–Sophia Children’s Hospital, Rotterdam, The Netherlands. y Princess Amalia Department of Pediatrics, Division of Neonatology, Isala Clinics, Zwolle, The Netherlands. x Department of Biostatistics, Erasmus University Medical Center, Rotterdam, The Netherlands.
Abstract: Short-term and long-term effects of neonatal pain and its analgesic treatment have been topics of translational research over the years. This study aimed to identify the long-term effects of continuous morphine infusion in the neonatal period on thermal pain sensitivity, the incidence of chronic pain, and neurological functioning. Eighty-nine of the 150 participants of a neonatal randomized controlled trial on continuous morphine infusion versus placebo during mechanical ventilation underwent quantitative sensory testing and neurological examination at the age of 8 or 9 years. Forty-three children from the morphine group and 46 children from the placebo group participated in this follow-up study. Thermal detection and pain thresholds were compared with data from 28 healthy controls. Multivariate analyses revealed no statistically significant differences in thermal detection thresholds and pain thresholds between the morphine and placebo groups. The incidence of chronic pain was comparable between both groups. The neurological examination was normal in 29 (76%) of the children in the morphine group and 25 (61%) of the children in the control group (P = .14). We found that neonatal continuous morphine infusion (10 mg/kg/h) has no adverse effects on thermal detection and pain thresholds, the incidence of chronic pain, or overall neurological functioning 8 to 9 years later. Perspective: This unique long-term follow-up study shows that neonatal continuous morphine infusion (10 mg/kg/h) has no long-term adverse effects on thermal detection and pain thresholds or overall neurological functioning. These findings will help clinicians to find the most adequate and safe analgesic dosing regimens for neonates and infants. ª 2015 by the American Pain Society Key words: Morphine, quantitative sensory testing, neonatal intensive care, follow-up, randomized controlled trial.
P
roviding adequate and evidence-based analgesia and sedation to neonates receiving intensive care is an ongoing challenge; one has to account for developmental changes in drug pharmacokinetics and
Received March 25, 2015; Revised May 29, 2015; Accepted June 15, 2015. No funding sources were provided. None of the authors report financial relationships that are relevant to this work. None of the authors report a conflict of interest. Supplementary data accompanying this article are available online at www.jpain.org and www.sciencedirect.com. Address reprint requests to Abraham J. Valkenburg, MD, PhD, Intensive Care and Department of Pediatric Surgery, Erasmus University Medical Center–Sophia Children’s Hospital, Dr. Molewaterplein 60, 3015 GJ Rotterdam, The Netherlands. E-mail: [email protected] 1526-5900/$36.00 ª 2015 by the American Pain Society http://dx.doi.org/10.1016/j.jpain.2015.06.007
926
pharmacodynamics as well as in the human nervous system.13,23 The repeated painful procedures in intensive care can lead to short-term hyperalgesia.22 Long-term follow-up of extremely preterm neonates showed a generalized decrease in thermal sensitivity, probably caused by tissue injury and modulation of nociceptor pathways.26 Therefore, analgesic treatment plays an important role in the neonatal intensive care unit (NICU). Opioids such as morphine are commonly used in the NICU and have both beneficial and adverse short-term effects. Morphine is an effective analgesic agent for neonates’ postoperative pain but not for acute procedural pain.1,4,5,20 However, continuous morphine infusion does not decrease the risk of poor neurological outcome after intensive care treatment.20
Valkenburg et al Animal studies on long-term cognitive functioning and neurodevelopmental outcome after neonatal morphine exposure found impaired learning abilities, neuronal degeneration,2 increased neuroapoptosis,3 and enhanced hippocampal gliosis (proliferation of damaged astrocytes) in rodents treated with morphine.14,21,25 In humans, participants of 2 randomized controlled trials (RCTs) on neonatal morphine administration were studied 5 years after the original RCT was conducted.1,20 The first, a small-scale follow-up of the NEOPAIN (Neurologic Outcomes and Pre-emptive Analgesia in Neonates) trial,9 showed that children who had received morphine (n = 14) had a smaller head circumference, weighed less, and had more social problems than children who had received placebo (n = 5). The second, performed in our institution, showed that children who had received morphine (n = 49) performed more poorly on 1 subtest of the intelligence scale than did the children who had received placebo (n = 41); other neurobehavioral outcomes and the incidences of chronic pain were comparable between the 2 groups.7 This unique cohort is being followed, and at the age of 8 years, participants were old enough for quantitative sensory testing (QST).6 Morphine is used worldwide for opioid analgesia in neonates, infants, and children; therefore, it is important to confirm that it does not have any negative long-term effects. In the current study, we evaluated thermal sensitivity, the occurrence of chronic pain, and neurological outcome in 8- to 9-year-old children who at neonatal age had received continuous morphine infusion. Our null hypothesis is that continuous morphine infusion at neonatal age has no adverse effects on thermal detection and pain thresholds, incidence of chronic pain, and neurological functioning at 8 to 9 years of age.
Methods Original Study Between 2000 and 2002, 150 term- and preterm-born neonates who received mechanical ventilation in 2 level III NICUs participated in a multicenter RCT. Morphine is one of the most frequently used analgesics across all age groups worldwide. At the time of the study,20 the debate was whether preterm newborns on ventilatory support should routinely receive a continuous morphine infusion. The aim of the RCT was to evaluate the effects of continuous intravenous morphine infusion on pain responses, incidence of intraventricular hemorrhage, and poor neurologic outcome (severe intraventricular hemorrhage, periventricular leukomalacia, or death). Seventy-three neonates were randomly assigned to the continuous morphine group (loading dose of 100 mg/kg followed by infusion of 10 mg/kg/h) and 77 to the placebo group (normal saline infusion). If pain or distress was noted, children in both groups received an open-label morphine bolus of 50 mg/kg and, if indicated, an openlabel morphine infusion (5–10 mg/kg/h) as rescue medication. Open-label morphine was administered to 27% of the children in the morphine group versus 40% of the
The Journal of Pain
927
placebo group (P = .10). Further details, including the background characteristics of the participants, can be found in the original article.20
Follow-Up Study (8–9 Years) The institutional ethics review boards of the 2 study sites (Erasmus University Medical Center–Sophia Children’s Hospital, Rotterdam, The Netherlands, and Isala Clinics, Zwolle, The Netherlands) approved the study plan. Parents of the 132 survivors were informed of the study and were asked for written informed consent. Seventeen participants from the morphine group were lost to follow-up and 5 parents refused to provide informed consent for the follow-up study. Sixteen patients of the placebo group were lost to follow-up and 5 parents refused informed consent for the follow-up study. The remaining 89 children and their parents were then invited for a follow-up visit in their hospital (either Rotterdam or Zwolle) (Fig 1). The nonparticipants had a lower birth weight and their first hospital stay was longer compared with the participants.8 Parents were asked to complete several questionnaires (see later discussion) before the visit. The visit consisted of 3 parts: QST by a trained researcher (A.J.V.), medical examination by a pediatrician (D.T.), and neuropsychological testing by a psychologist (M.vD.). These health professionals were blinded to the participants’ study condition (continuous morphine infusion versus placebo) in the original RCT.
QST Participants underwent QST (see Supplementary Material) in a quiet hospital room, with a stable room temperature (20–22 C). Parents were present in the room and were instructed not to interfere during the test. Reaction time was measured using the baseline speed task for the dominant hand (Amsterdam Neuropsychological Tasks, Version 3.1; Boom Test Publishers, Amsterdam, The Netherlands).
Reference Data QST (Control Group) We compared the detection and pain thresholds of both the placebo and the morphine group with detection and pain thresholds of 28 healthy control individuals (aged 8–10 years). The control values were collected by our research group and obtained using the same protocol as for QST.24 Healthy term-born controls without a history of intensive care admission or severe early pain were recruited. Participants of other studies were asked whether they could recommend someone in the age range of 8– 18 years who would be interested in volunteering. Furthermore, children were recruited at primary schools in Rotterdam. Eventually, we selected the children aged 8 to 10 years as the reference group for the present study. Control individuals were included at Erasmus Medical Center in Rotterdam from October 2011 to March 2013. The local institutional review board approved this study. Informed consent from the parents and assent of the children was obtained before participation.
928
Long-Term Effects of Neonatal Morphine Infusion
The Journal of Pain
Figure 1. Flowchart based on the CONSORT checklist.
Chronic Pain Questionnaire
Statistical Analysis
Parents of all children also completed the Chronic Pain Questionnaire (Dutch version)18; this questionnaire examines the incidence and characteristics of pain in the 3 months before the visit (location, frequency, duration, and intensity). Chronic pain is defined as a pain episode with duration of 3 months or longer (either recurrent pain with pain-free intervals or continuous pain). The questionnaire was developed and validated for children between 4 and 18 years old and can be completed in less than 10 minutes.
Summary statistics are presented as medians (interquartile range [IQR]) for continuous variables and as percentages for categorical variables. Data were compared between the continuous morphine group and the placebo group using the Mann-Whitney test for continuous, nonnormal data and the c2 test (or Fisher exact test in the case of low predicted cell counts) for nominal data. IQ data (Wechsler Intelligence Scale for Children) are presented as means (standard deviation [SD]) and compared between the morphine and placebo groups using the independent samples t-test. The QST data of the morphine and placebo groups were also compared with data of 28 healthy control individuals using analysis of covariance, correcting for age using a post hoc test with Bonferroni correction for multiple comparisons. For the multivariate analysis of the QST data, we used robust regression analysis with MM estimation28 to account for the fact that the model residuals were not normally distributed. The weight function was the Tukey bisquare estimator. A model was built for each of the 6 QST modalities (see Supplementary Material), with treatment condition (morphine vs placebo), total amount of morphine in the first 28 days of life, age, sex, study site, and propensity score as predictor variables. The propensity score was derived from logistic regression analysis in which the treatment condition (morphine or placebo) was the criterion and the covariables were study site, sex, gestational age, birth weight, Clinical Risk Index for Babies, and socioeconomic status. In this way, several
Medical Examination All children were examined by a pediatrician (D.T.) who referred them for further diagnosis or treatment if indicated. Weight, height, and head circumference were measured and plotted against sex-matched reference values for The Netherlands (fourth nationwide growth study, 1997).10 The neurological examination was based on the Touwen assessment of minor neurological dysfunctions.15 We assessed 41 items in 5 domains (posture/muscle tone, reflexes, involuntary movements, coordination/balance, and cranial nerve dysfunctions). Minor neurological dysfunctions were defined as the presence of 2 or more deviant items in at least 1 domain.
Neuropsychological Testing A trained psychologist (M.vD.) tested the participants’ IQ with the Wechsler Intelligence Scale for Children-III (Dutch version).16
Valkenburg et al
The Journal of Pain
relevant covariables are combined into 1 weighted score.7,8,27 Logistic regression analyses were applied with prevalence of chronic pain and minor neurological dysfunctions (0 or 1 vs >1) as dichotomous outcome variables and with the treatment condition (continuous morphine vs placebo) and the total amount of morphine in the first 28 days after birth as predictor variables. P values (2-sided) of less than .05 were considered statistically significant. Data were analyzed with SPSS software, version 19.1 (IBM Corp, Armonk, NY) and SAS, version 9.3 (SAS Institute Inc, Cary, NC).
Results Background Characteristics Of the 132 survivors, 43 children from the morphine group and 46 children from the placebo group participated in this follow-up study (Fig 1). Background characteristics were comparable between the 2 groups (Table 1). Total morphine exposure in the first 28 days of life was statistically significantly higher in the morphine group (P < .001) than in the placebo group. Twelve children (26.1%) in the placebo group did not receive any opioids. Mean IQ was 101 (SD = 18) for the placebo group versus 99 (SD = 19) for the morphine group (P = .63). Twenty-eight controls were included to compare thermal and pain sensitivity as well as incidence of chronic pain among the morphine group, placebo group, and healthy controls.
QST QST was performed in 41 children (89%) of the placebo group, 37 children (86%) of the morphine group, and 28 healthy control individuals. Nine children did not attend
Table 1.
929
the follow-up visit (however, parents did complete the questionnaires), and the equipment was not available on the remaining 2 occasions. QST was feasible in all participants with regard to understanding the instructions and completing the test. No significant differences in detection or pain thresholds were found among the placebo group, morphine group, and the control group using both the method of limits and the method of levels (Table 2). Ten (24%) of the children in the placebo group did not establish a pain threshold for cold before the minimum temperature (0 C) was reached versus 9 (25%) of the morphine group and 16 (57%) of the control individuals (P = .006; post hoc test: significant difference between both placebo group and morphine group versus controls). Ten (24%) of the children in the placebo group did not establish a pain threshold for heat before the maximum temperature (50.0 C) was reached, versus 8 (22%) of the children in the morphine group and 13 (46.4%) of the control individuals (P = .38). Regression estimates for the 6 QST modalities are presented in Tables 3–5. Neither treatment condition (morphine group vs placebo) nor the total amount of morphine in the first 28 days was a statistically significant predictor in any of the models. For the cold and warmth detection thresholds (method of limits), IQ was a statistically significant covariate (P = .03 and P = .01, respectively). The higher the IQ, the more sensitive the children were for the detection of cold and warmth (Table 3). Children tested in the hospital in Zwolle (n = 39) had statistically significant lower mean cold pain thresholds (P < .001) and borderline statistically significant lower mean heat pain thresholds (P = .054) (Table 4). For the warm detection thresholds (method of levels), IQ was a statistically significant covariate (P = .007).
Background Characteristics (Distinguished by Group)
Characteristics Male sex Birth characteristics Gestational age, wk Birth weight, g Age, y Height, cm SDS height for agey Weight, kg SDS weight for agey Head circumference, cm SDS head circumference for agey IQz Reaction time, msx Test location Rotterdam Zwolle
PLACEBO GROUP (N = 46) 28 (61%) 31 (28–32) 1520 (995–2145) 8.9 (8.7–9.3) 137 (131–140) .02 (.7 to 1.1) 29.2 (26.5–32.8) .05 (.6 to .7) 52 (51–53) .20 (.8 to .7) 101 (18) 391 (354–442) 25 (54%) 21 (46%)
Abbreviation: SDS, standard deviation score. NOTE. Values are median (IQR) or number (%). *Comparison Mann-Whitney test for continuous variables and Fisher exact test for categorical variables. yAccording to reference values (The Netherlands, 1997). zIQ (Wechsler Intelligence Scale for Children), mean (SD), P value from t-test. xAmsterdam Neuropsychological Tasks.
MORPHINE GROUP (N = 43) 28 (65%) 30 (29–32) 1400 (1024–1788) 8.9 (8.8–9.1) 135 (131–140) .04 (.5 to .7) 27.9 (24.6–31.6) .01 (1.3 to .8) 52 (51–54) .03 (1.0 to .6) 99 (19) 390 (331–442) 25 (58%) 18 (42%)
P VALUE* .83 .39 .70 .54 .42 .72 .29 .53 .90 .92 .63 .82 .83
930
Long-Term Effects of Neonatal Morphine Infusion
The Journal of Pain
Table 2.
Results of QST, Distinguished by Group Modality
Method of limits Cold detection threshold, C Warm detection threshold, C Cold pain threshold, C Threshold not reached at 0 Cy Heat pain threshold, C Threshold not reached at 50 Cy Method of levels Cold detection threshold, C Number of stimuli Warm detection threshold, C Number of stimuli
PLACEBO GROUP (N = 41)
MORPHINE GROUP (N = 37)
CONTROL GROUP (N = 28)
P VALUE*
29.6 (2.4) 34.8 (1.4) 11.8 (8.4) 10 (24%) 45.0 (3.5) 10 (24%)
29.9 (1.4) 34.5 (1.4) 12.3 (8.9) 9 (25%) 44.9 (4.3) 8 (22%)
30.5 (1.0) 34.2 (1.5) 9.3 (10.1) 16 (57%) 45.2 (5.0) 13 (46%)
.15 .23 .40 .006z .74 .38
30.7 (1.2) 11 (3) 33.3 (1.1) 11 (4)
31.1 (.7) 10 (2) 33.2 (1.1) 10 (2)
30.5 (1.5) 11 (3) 33.6 (.9) 10 (3)
.27 .38 .84 .30
NOTE. Values are mean (SD) or number (%). *Analysis of covariance test for the comparison among the 3 groups corrected for age. yPatients in whom 1 or more times the pain threshold was not reached (the child did not press the button before the temperature of the thermode reached its minimum/ maximum (.0 C and 50.0 C, respectively). zPost hoc Bonferroni correction: placebo group and morphine group versus control individuals, P < .05.
Chronic Pain Chronic pain (duration longer than 3 months) was reported in 4 (9%) children in the placebo group versus 5 (12%) in the morphine group (P = .42). Abdominal pain was the most common type of pain; it was present in 3 of the children in the placebo group and 3 children in the morphine group (P = 1.00). The second most common type of pain was headache; it was present in 4 of the children in the placebo group versus 2 in the
Robust Regression Estimates of Intervention, With Adjustment for Different Covariates: Detection Thresholds (Method of Limits)
Table 3.
Detection Threshold
ESTIMATE
Detection threshold for cold Intercept 31.71 Treatment condition* .14 Total morphiney .08 Sex .18 Age .31 Propensity scorez .05 Study site .03 IQx .01 Detection threshold for warmth Intercept 32.28 Treatment condition* .35 Total morphiney .12 Sex .38 Age .45 Propensity scorez .32 Study site .47 IQx .02
95% CI LIMITS
P VALUE
26.04 to 37.39 .32 to .60 .12 to .29 .62 to .26 .93 to .31 .53 to .63 .40 to .46 .001 to .02
<.001 .55 .43 .43 .32 .86 .88 .03
23.75 to 40.80 1.01 to .31 .43 to .20 .27 to 1.03 .48 to 1.39 1.15 to .50 .15 to 1.11 .039 to .004
<.001 .30 .47 .25 .34 .44 .14 .01
Abbreviation: CI, confidence interval. *Morphine group versus placebo group. yTotal morphine in first 28 d. zPropensity score based on neonatal characteristics (study location, sex, duration mechanical ventilation, gestational age, Clinical Risk Index for Babies, and socioeconomic status [middle and high vs low]). xIQ (Wechsler Intelligence Scale for Children).
morphine group (P = .17). The intensity of the chronic pain was low; the median (IQR) visual analog scale score was 1.00 (.50–2.05) for the morphine group versus 1.85 (.63–4.20) in the placebo group (P = .41). Two children in the morphine group complained of pain 2 or 3 times a month, and 3 children complained 2 or more times a week. Two children in the placebo group complained of pain once a month, and 2 children complained 2 or more times a week (P = .91). In the control group, chronic pain was reported in 5 (18%) of the children.
Robust Regression Estimates of Intervention, With Adjustment for Different Covariates: Pain Thresholds (Method of Limits)
Table 4.
Pain Threshold Pain threshold for cold Intercept Treatment condition* Total morphiney Sex Age Propensity scorez Study site IQx Pain threshold for warmth Intercept Treatment condition* Total morphiney Sex Age Propensity scorez Study site IQx
ESTIMATE
95% CI LIMITS
P VALUE
.95 1.91 .25 .51 1.44 .24 7.78 .02
57.77 to 59.66 2.80 to 6.61 2.05 to 2.54 5.11 to 4.09 4.97 to 7.85 5.89 to 5.41 12.22 to 3.35 .10 to .14
.97 .43 .83 .83 .66 .93 <.001 .77
46.71 .32 .05 1.16 .22 .52 1.94 .04
20.47 to 72.94 1.79 to 2.42 1.08 to .99 3.20 to .88 2.65 to 3.10 3.06 to 2.02 .04 to 3.91 .09 to .01
<.001 .77 .93 .26 .88 .69 .05 .15
Abbreviation: CI, confidence interval. *Morphine group versus placebo group. yTotal morphine in first 28 d. zPropensity score based on neonatal characteristics (study location, sex, duration mechanical ventilation, gestational age, Clinical Risk Index for Babies, and socioeconomic status [middle and high vs low]). xIQ (Wechsler Intelligence Scale for Children).
Valkenburg et al
The Journal of Pain
Robust Regression Estimates of Intervention, With Adjustment for Different Covariates: Detection Thresholds (Method of Levels)
Table 5.
Detection Threshold
ESTIMATE
Detection threshold for cold Intercept 34.66 Treatment condition* .31 Total morphiney .03 Sex .05 Age .48 Propensity scorez .27 Study site .13 IQx .004 Detection threshold for warmth Intercept 31.01 Treatment condition* .18 Total morphiney .07 Sex .17 Age .38 Propensity scorez .42 Study site .20 IQx .01
95% CI LIMITS
P VALUE
28.74 to 40.59 .13 to .75 .22 to .17 .46 to .36 1.12 to .16 .79 to .25 .28 to .54 .007 to .02
<.001 .16 .80 .81 .14 .31 .53 .46
26.45 to 35.57 .54 to .19 .10 to .23 .18 to .51 .11 to .87 .04 to .89 .53 to .14 .02 to .003
<.001 .35 .44 .35 .13 .07 .25 .007
Abbreviation: CI, confidence interval. *Morphine group versus placebo group. yTotal morphine in first 28 d. zPropensity score based on neonatal characteristics (study location, sex, duration mechanical ventilation, gestational age, Clinical Risk Index for Babies, and socioeconomic status [middle and high vs low]). xIQ (Wechsler Intelligence Scale for Children).
Logistic regression analysis revealed that neither treatment condition (morphine vs placebo) nor the total amount of morphine in the first 28 days was a significant predictor for the presence of chronic pain (P = .68 vs P = .84).
Neurological Examination Two children in the placebo group and 2 children in the morphine group were severely intellectually and developmentally disabled and did therefore not attend the follow-up visit. Neurological examination was performed in 38 children of the morphine group and 41 of the placebo group (Table 6). The neurological examination was normal in 29 (76%) of the children in the morphine group and 25 (61%) of the children in the control group (P = .14). Logistic regression revealed that neither group (morphine vs placebo) nor the total Table 6.
931
amount of morphine in the first 28 days was a significant predictor for the presence of minor neurological dysfunctions (P = .30 vs P = .23). The presence of minor neurological dysfunctions at 8 to 9 years of age was not related to the presence of intraventricular hemorrhages in the neonatal period (P = .26).
Discussion A cohort of children who as neonates participated in an RCT on continuous morphine infusion versus placebo for mechanical ventilation was studied at the age of 8 to 9 years. In the multivariate analyses, we found that the treatment condition (morphine vs placebo) was not a significant predictor in any of the 6 QST modalities. The multivariate analysis did show correlations between IQ and the detection thresholds, with both the reaction time–dependent method of limits and the reaction time–independent method of levels. However, all correlations were weak and clinically not significant. In another study by our group, we found that in 5-yearold children, detection threshold levels assessed with the method of limits were significantly correlated with IQ (r = .64; P = .006), whereas the detection threshold levels assessed with the method of levels were not correlated with IQ.6 However, in the present study, we also found a weak correlation between IQ and the thermal detection thresholds assessed with the reaction time–independent method of levels. This suggests that children with a higher IQ could have a better attentiveness as well.
Long-Term Effects of Surgery and Morphine Hermann et al12 reported long-term hypoalgesia for heat pain in both preterm and term-born 9- to 14-year-old children who had received neonatal intensive care. The preterm neonates in that study underwent a mean number of 172 invasive procedures in the first week of life; however, no more than 53% of them received analgesics. The hypoalgesia in former extremely preterm-born 11-year-olds, reported by Walker et al,26 was most marked in those who had undergone neonatal surgery. In contrast, another study on long-term effects of neonatal surgery (all neonates received opioids postoperatively) showed no differences in the cold and warmth
Minor Neurological Dysfunctions (Distinguished by Group) Dysfunction
PLACEBO GROUP (N = 41)
MORPHINE GROUP (N = 38)
P VALUE*
Minor neurological dysfunctions Mild deviations in posture/muscle tone Mild deviations in reflexes Presence of involuntary movements Mild deviations in coordination/balance Mild deviations in cranial nerve function
16 (39) 6 (15) 4 (10) 0 (0) 5 (12) 1 (2)
9 (24) 2 (5) 1 (3) 0 (0) 7 (18) 0 (0)
.14 .27 .36 1.00 .53 1.00
NOTE. Values are number (%). *Fisher exact test.
932
The Journal of Pain
detection thresholds between the neonatal surgery group and the control group at the age of 9 to 12 years.19 Combining the previous data with the results of the present study, we hypothesize that neonatal injury or surgery is likely to have more pronounced long-term effects on pain processing than neonatal morphine treatment itself.
Chronic Pain The prevalence of chronic pain was comparable between the 2 groups and lower than the reference values (ie, 11% for boys vs 13% for girls).18 The prevalence of chronic pain in both groups had decreased since the 5-year follow-up visit, ie, from 14% to 9% in the placebo group and from 15% to 12% in the morphine group.7
Neurological Functioning The results of the neurological examination were normal for most of the children. The prevalence of minor neurological dysfunctions (ie, 39% of the placebo group vs 24% of the morphine group) is comparable with the prevalence in a reference cohort of term-born children (ie, 50% prevalence of minor neurological dysfunctions).15 Mild deviations in coordination/balance control (a minor neurological dysfunction that could indicate cerebellar dysfunction) were more common in the morphine group in the present study. Research in rodents suggests that morphine treatment has a negative effect on the development of cerebellar neurons.11 A study in human preterm-born infants17 found that they were more at risk for cerebellar injuries during the neonatal period than were term-born infants and that at adolescent age their cerebellum volume was smaller than that in term-born adolescents. The minor neurological dysfunctions observed may reflect the combined consequences of prematurity and critical illness along with potential effects of morphine, not the effects of morphine in isolation. The pilot follow-up study of the NEOPAIN trial9 found an overall lower prevalence of neurological soft signs (ie, 20% in the placebo group [n = 5] vs 14% in the morphine group [n = 14]); however, the researchers did not provide details on which neurological soft signs (comparable with minor neurological dysfunctions) were assessed. In that study, the children
References 1. Anand KJ, Hall RW, Desai N, Shephard B, Bergqvist LL, Young TE, Boyle EM, Carbajal R, Bhutani VK, Moore MB, Kronsberg SS, Barton BA: Effects of morphine analgesia in ventilated preterm neonates: primary outcomes from the NEOPAIN randomised trial. Lancet 363:1673-1682, 2004 2. Atici S, Cinel L, Cinel I, Doruk N, Aktekin M, Akca A, Camdeviren H, Oral U: Opioid neurotoxicity: comparison of morphine and tramadol in an experimental rat model. Int J Neurosci 114:1001-1011, 2004 3. Bajic D, Proudfit HK: Projections from the rat cuneiform nucleus to the A7, A6 (locus coeruleus), and A5 pontine noradrenergic cell groups. J Chem Neuroanat 50-51:11-20, 2013
Long-Term Effects of Neonatal Morphine Infusion in the morphine group had 7% smaller head circumference and 4% less body weight than children in the placebo group. Although several children in our study were born prematurely or were born small for gestational age, most children in both groups now had normal height, weight, and head circumference for their age. The preterm neonates in the NEOPAIN trial received higher doses of morphine, up to 30 mg/kg/h of morphine.1 The children in the morphine group in our study received 10 mg/kg/h of morphine at neonatal age; in the case of pain or distress, children in both groups received open-label morphine as rescue medication.
Limitations Given the range of variation in data on pain thresholds in other studies4 and the fact that the study site (Zwolle vs Rotterdam) was a significant covariate for the cold pain thresholds in the present study, we suggest that assessing pain thresholds is sensitive to variations in instructions and methodology. However, this does not influence the comparison among the morphine group, placebo group, and control group in the present study. Children in all the study groups received the same instructions.
Conclusions We found in the present study that continuous morphine infusion (10 mg/kg/h) during the neonatal period had no adverse effects on thermal detection and pain thresholds, the incidence of chronic pain, or overall neurological functioning 8 to 9 years later. Clinicians around the world encounter postoperative and procedural pain in neonates on a day-to-day basis. These are important findings in view of the concerns on neurotoxicity of opioids and anesthetic agents and will help clinicians to find the most adequate and safe analgesic dosing regimens for neonates and infants.
Supplementary Data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jpain.2015.06.007.
4. Bellu’ R, Dewaal K, Zanini R: Opioids for neonates receiving mechanical ventilation. a systematic review and meta-analysis. Arch Dis Child-Fetal 95:241-251, 2009 5. Carbajal R, Lenclen R, Jugie M, Paupe A, Barton BA, Anand KJS: Morphine does not provide adequate analgesia for acute procedural pain among preterm neonates. Pediatrics 115:1494-1500, 2005 6. de Graaf J, Valkenburg AJ, Tibboel D, van Dijk M: Thermal detection thresholds in 5-year-old preterm born children. Early Hum Dev 88:487-491, 2012 7. de Graaf J, van Lingen RA, Simons SHP, Anand KJS, Duivenvoorden HJ, Weisglas-Kuperus N, Roofthooft DWE, Jebbink LJMG, Veenstra RR, Tibboel D, van Dijk M: Longterm effects of routine morphine infusion in mechanically ventilated neonates on children’s functioning: five-year
Valkenburg et al
The Journal of Pain
933
follow-up of a randomized controlled trial. Pain 152: 1391-1397, 2011
Wouden JC: Pain in children and adolescents: a common experience. Pain 87:51-58, 2000
8. de Graaf J, van Lingen RA, Valkenburg AJ, WeisglasKuperus N, Groot Jebbink L, Wijnberg-Williams B, Anand KJ, Tibboel D, van Dijk M: Does neonatal morphine use affect neuropsychological outcomes at 8 to 9 years of age? Pain 154:449-458, 2013
19. Schmelzle-Lubiecki BM, Campbell KA, Howard RH, Franck L, Fitzgerald M: Long-term consequences of early infant injury and trauma upon somatosensory processing. Eur J Pain 11:799-809, 2007
9. Ferguson SA, Ward WL, Paule MG, Hall RW, Anand KJ: A pilot study of preemptive morphine analgesia in preterm neonates: effects on head circumference, social behavior, and response latencies in early childhood. Neurotoxicol Teratol 34:47-55, 2012 10. Fredriks AM, van Buuren S, Wit JM, VerlooveVanhorick SP: Body index measurements in 1996-7 compared with 1980. Arch Dis Child 82:107-112, 2000 11. Hauser KF, Gurwell JA, Turbek CS: Morphine inhibits Purkinje cell survival and dendritic differentiation in organotypic cultures of the mouse cerebellum. Exp Neurol 130:95-105, 1994 12. Hermann C, Hohmeister J, Demirakc¸a S, Zohsel K, Flor H: Long-term alteration of pain sensitivity in schoolaged children with early pain experiences. Pain 125: 278-285, 2006 13. Johnston CC, Fernandes AM, Campbell-Yeo M: Pain in neonates is different. Pain 152:S65-S73, 2011 14. Juul SE, Beyer RP, Bammler TK, Farin FM, Gleason CA: Effects of neonatal stress and morphine on murine hippocampal gene expression. Pediatr Res 69:285-292, 2010 15. Kikkert HK, De Jong C, Hadders-Algra M: Minor neurological dysfunction and IQ in 9-year-old children born at term. Dev Disabil Res Rev 53:e16-25, 2011 16. Kort W, Schittekatte M, Dekker PH, Verhaeghe P, Compaan EL, Bosmans M, Vermeir G: WISC-III NL Wechsler Intelligence Scale for Children. In: Derde Editie NL (ed): Handleiding en Verantwoording. Amsterdam, Harcourt Test Publishers/Nederlands Instituut voor Psychologen, 2005
20. Simons SH, van Dijk M, van Lingen RA, Roofthooft D, Duivenvoorden HJ, Jongeneel N, Bunkers C, Smink E, Anand KJS, van den Anker JN, Tibboel D: Routine morphine infusion in preterm newborns who received ventilatory support: a randomized controlled trial. JAMA 290:2419-2427, 2003 21. Sternberg WF, Scorr L, Smith LD, Ridgway CG, Stout M: Long-term effects of neonatal surgery on adulthood pain behavior. Pain 113:347-353, 2005 22. Taddio A, Shah V, Atenafu E, Katz J: Influence of repeated painful procedures and sucrose analgesia on the development of hyperalgesia in newborn infants. Pain 144:43-48, 2009 23. Tibboel D, Anand KJS, van den Anker JN: The pharmacological treatment of neonatal pain. Semin Fetal Neonat Med 10:195-205, 2005 24. van den Bosch GE, White T, Tibboel D, van Dijk M: Functional MRI pain studies in children? Yes, we (s)can!. Pediatr Radiol 43:1235-1236, 2013 25. Vien TN, Gleason CA, Hays SL, McPherson RJ, Chavkin C, Juul SE: Effects of neonatal stress and morphine on kappa opioid receptor signaling. Neonatology 96:235-243, 2009 26. Walker SM, Franck LS, Fitzgerald M, Myles J, Stocks J, Marlow N: Long-term impact of neonatal intensive care and surgery on somatosensory perception in children born extremely preterm. Pain 141:79-87, 2009
17. Manto MU, Jissendi P: Cerebellum: links between development, developmental disorders and motor learning. Front Neuroanat 6:1-10, 2012
27. Wright AA, Zhang B, Ray A, Mack JW, Trice E, Balboni T, Mitchell SL, Jackson VA, Block SD, Maciejewski PK, Prigerson HG: Associations between end-of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment. JAMA 300:1665-1673, 2008
18. Perquin CW, Hazebroek-Kampschreur AA, Hunfeld JA, Bohnen AM, van Suijlekom-Smit LW, Passchier J, van der
28. Yohai VJ, Zamar RH: Optimal locally robust M-estimates of regression. J Stat Plan Inference 64:309-323, 1997