Do We Still Hurt Newborn Babies?: Title and subTitle BreakA Prospective Study of Procedural Pain and Analgesia in Neonates FREE

ADVANCES IN perinatal care have increased the survival of preterm neonates in neonatal intensive care units (NICUs) worldwide. The physiological instability and underlying diseases of these infants necessitate various invasive procedures, such as endotracheal intubation, heelsticks, insertion of indwelling venous and arterial catheters, as well as nasal, tracheal, and gastric suctioning. Although previous studies^{1 - 4} have reported the frequency of daily procedures in the NICU (Table 1), many procedures such as nasopharyngeal tube insertions and nasal suctioning were not included, and the number of failed procedures has never been evaluated, to our knowledge.

Despite the current knowledge that low birth weight infants are able to experience pain, many daily procedures are still performed without pharmacological or nonpharmacological analgesic therapy.^{2 ,5 - 9} Indeed, preterm neonates are highly sensitive to pain^{10 - 12} and amass acute responses to painful procedures.^{13 - 14} Short- and long-term effects of painful procedures in neonates^{15 - 16} occur as a consequence of their immature and vulnerable nervous systems, reflected in an altered pain response,¹⁷ possibly leading to changes in neural development. Studies in animals indicate developmental changes in the brain^{18 - 19} and in the spinal dorsal horn²⁰ that are associated with neonatal pain. A higher frequency of painful invasive procedures in low birth weight infants has been associated with a greater pain response at 32 weeks compared with controls.²¹ Although it is unknown whether these changes in pain response persist until older age,^{22 - 23} more cognitive²⁴ and psychopathological^{25 - 32} problems have been reported in children who are born preterm.^{33 - 34}

A growing awareness in the past several years of the importance of adequate neonatal pain management led to the discussion whether analgesia should be given to all ventilated newborns.^{35 - 36} Although there is controversy about the risks and benefits of continuous opioid administration (and as a consequence neonatologists are reluctant to prescribe them), recently published international consensus statements^{37 - 38} have provided neonatal pain management guidelines, including those for procedural interventions. Therefore, we hypothesized that analgesic therapy is frequently used and that procedural pain is minimal in neonates admitted to the NICU. Prospectively, we collected bedside data on the number and type of daily painful procedures performed in 151 neonates during their stay in a tertiary care NICU, including failed procedures (eg, multiple attempts to insert peripheral venous catheters). These data were integrated with the results of a questionnaire evaluating the opinions of experienced clinicians about the pain of different procedures. Furthermore, the analgesic therapy used was evaluated and compared with current guidelines.

A pain research team (including a neonatologist [D.R.], pediatric intensivist [D.T.], research nurse, and psychologist [M. van D.]) with extensive research^{39 - 50} and clinical experience compiled a checklist containing all daily painful procedures, invasive and noninvasive, including multiple attempts. The checklist designed for this study was tested and refined during a pilot study in December 2000 and January 2001, and 34 procedures were included in the final version.

From February 1, 2001, to July 31, 2001, we documented all procedures performed in all neonates during the first 14 days of admission in a tertiary care NICU. The unit is part of the biggest children's hospital in the Netherlands, which comprises a perinatal center and pediatric surgery setting. Patients older than 3 days at admission and infants discharged or transferred within 24 hours after admission were excluded from the study. Nurses and physicians noted all procedures performed in real time each day, including the number of attempts for each procedure. The recorded procedures were cross-checked by the researchers every day to ensure accuracy. Background variables and mode of respiratory support were noted daily, and the Clinical Risk Index for Babies (CRIB)⁵¹ was scored as a measure of severity of illness.

In addition, all administered analgesics were noted during the study days. Standardized pharmacological analgesic therapy for procedural pain in our NICU is limited to a morphine loading dose (100 µg/kg of body weight) before tracheal intubation (unless intubation occurs in the delivery room) and before the insertion of thoracostomy tubes. Infants with indwelling thoracostomy tubes receive a continuous morphine infusion (10 µg/kg per hour) until the tube is removed.

To estimate the pain of NICU procedures, we developed a questionnaire listing all invasive procedures from our checklist, with 2 noninvasive procedures (diaper change and cranial ultrasound) included as control variables. The questionnaire was distributed among the nurses and physicians of (1) the NICU, where the procedures were also counted, (2) a pediatric surgical intensive care unit present in the same hospital where newborns with major congenital anomalies are admitted, and (3) the level III NICU of another, nonuniversity, hospital (Isala Clinics). Participants were asked to estimate a rating from 0 (not painful) to 10 (most painful) for each procedure, without taking specific circumstances into account.

Procedures were counted per calendar day. Because the first and last study days are usually shorter than 24 hours, the numbers of procedures were corrected for the actual length of stay on these days.

Random regression modeling (PROC MIXED; SAS Institute, Cary, NC) was used to simultaneously estimate the effect of the time-varying covariates respiratory support (no support, nasal oxygen, continuous positive airway pressure, and mechanical ventilation), postnatal age, gestational age, and length of study on the number of procedures performed. Because a procedural volume difference on the first study day compared with the other days was expected, the study days were dichotomized into 2 variables: 0 (1 day) and 1 (2-14 days). The outcome variable, ie, frequency of procedures, was log-transformed (base 10) to achieve a normal distribution. The model incorporated random intercepts and random slopes.

Multiple regression analysis was performed to estimate the effect of background variables (profession, sex, age, unit, hospital [Erasmus MC–Sophia Children's Hospital or Isala Clinics], parent [yes or no], and years of NICU experience) on the pain scoring of the participants.

Data are presented as mean ± SD.

One hundred fifty-one neonates were included in the study; 89 other newborns who were discharged within 24 hours were excluded. Table 2 lists the study subjects' background characteristics and primary diagnoses. Most neonates (n = 129) were admitted and enrolled on the first postnatal day. Gestational ages ranged from 25.3 to 42.0 weeks, with a mean age of 32.4 ± 4.5 weeks. The CRIB⁵¹ scores ranged from 0 to 16, with a mean score of 3.9 ± 3.3. Study subjects required respiratory support consisting of mechanical ventilation on 49.6% of study days (one third high-frequency oscillation and two thirds conventional ventilation), continuous positive airway pressure on 22.5% of study days, nasal oxygen on 15.6% of study days, and subjects required no respiratory support on 12.2% of study days. During 55.2% of the study days, patients had an arterial line (42.8% peripheral arterial lines and 12.3% umbilical arterial lines).

During 1375 patient-days, 19 674 procedures were performed, with a mean NICU stay of 9.1 ± 4.4 days per patient. Table 3 gives the procedures rank-ordered by their frequency. Suctioning of nasal, endotracheal, and nasopharyngeal tubes constituted 63.6% of the performed procedures. The mean number of procedures per neonate per day was 14.3 ± 4, with a range of 0 to 53 procedures per day. Almost one third (30.9%) of the 1076 insertions of intravenous cannulae were not successful. Procedures for placement of central venous catheters, peripheral arterial catheters, and umbilical catheters were not successful in 45.6%, 37.5%, and 34.6% of attempts, respectively. Failure rates for venipunctures and lumbar punctures were 21.0% and 17.5%, respectively (Figure 1).

Random regression modeling (Table 4) showed significantly higher frequencies of procedures during the first study day compared with days 2 to 14 (P<.001). The frequency of procedures was not predicted by gestational age (P = .51), day of admission (P = .50), or postnatal age (P = .72).

Procedures were performed with significantly higher frequencies in patients receiving nasal oxygen, continuous positive airway pressure, and ventilation (P<.001 for all) compared with those without respiratory support.

Two hundred forty-seven questionnaires were distributed, with a response rate of 59.9% (n = 148), which was similar for nurses and physicians (Table 5 gives their background characteristics). The mean pain score, across all respondents and all procedures, on the 10-point scale was 5.2 ± 1.3. The mean pain score per procedure varied from 1.7 ± 1.6 for diaper change to 8.9 ± 1.4 for intubation. Pain ratings given by the nurses and physicians are shown in Figure 2.

Because procedures were scored on a 10-point scale, results of this questionnaire can be compared with a frequently used pain score, the visual analog scale. A visual analog scale score above 4 is generally used as a criterion for administration of extra analgesic therapy.⁴⁵ Therefore, procedures with pain scores above 4 can be considered as moderately painful. Procedures that received lower scores included the control variables, diaper change and cranial ultrasound, as well as insertion of nasal cannulae, X-rays, and the removal of nasogastric tubes. All other procedures were considered moderately to severely painful.

Multiple regression analyses (Table 6) showed that nurses scored various procedures as more painful than physicians (β = .30, P = .001), caregivers who were parents themselves scored procedures lower than persons without children (β = .27, P = .002), and ratings between the 2 hospitals were somewhat different (β = .20, P = .03). The multiple regression coefficient (R = 0.43) explained about 15% of the variance in the model.

Pharmacological analgesic treatment was studied in 126 patients (Figure 3); 25 neonates were excluded because they were enrolled in a double-blind, randomized controlled trial comparing morphine with placebo. Fifty patients (39.7%) did not receive any analgesic therapy during the study. Analgesic treatment regimens consisted of a morphine loading dose (given before intubation in 67.8% of administrations), continuous morphine infusion, or morphine in combination with vecuronium bromide, midazolam hydrochloride, or acetaminophen supplements. Fewer than 35% of newborns received analgesic therapy each day. Most therapy was administered during the first day, when 41 infants (32.5%) received analgesics, and decreased within 2 weeks to 12.2% and 14.6% of infants receiving analgesics on days 13 and 14, respectively.

Analgesic therapy, pharmacological and nonpharmacological, was rarely applied before invasive procedures. Analgesics were generally used only before tracheal intubation or insertion of a chest tube and usually were not given in association with other procedures with high pain scores. Although some nurses used pacifiers and tried to comfort infants during and after procedures by handling, nonpharmacological analgesic treatment was not given routinely for any of the procedures.

In sharp contrast to the accumulating evidence that repetitive pain is harmful in newborns, and despite major clinical advances over the past 10 years, neonates experience up to 14 painful procedures per day, and, remarkably, more than 65% of the patients in this study did not receive appropriate analgesic therapy.

We report a higher number of painful procedures in this study compared with previous studies, perhaps because we used a more extensive list of procedures and accounted for failed procedures (Table 1). For instance, Benis and Suresh³ reported a mean of 6 procedures per day during the entire NICU stay of 15 neonates. Barker and Rutter¹ reported an increased number of procedures in infants younger than 31 weeks' gestation compared with older infants, but did not mention the length of NICU stay, which may explain the exposure to a greater number of invasive procedures. In the present study, however, the frequency of procedures per day was not related to gestational age, perhaps because term neonates admitted to the NICU may have a severity of illness comparable to that of preterm neonates.

To our knowledge, we are the first to report prospective data on a substantial number of failed procedures in the NICU. Although attempts by even the most experienced clinicians may be unsuccessful, the relative inexperience of trainees may partly explain the high proportion of failed procedures in this study. These findings may, therefore, be applicable to NICUs located in other academic centers.

Efforts should be aimed at minimizing the number of invasive procedures, as stated in recent consensus guidelines.^{37 - 38} Our study showed that the number of procedures is significantly higher in neonates requiring nasal oxygen, continuous positive airway pressure, or ventilation and that procedures occur mostly on the first day of admission, because of initial stabilization, monitoring, and diagnostic evaluation. Furthermore, the number of heelsticks and venipunctures is subsequently decreased in patients having arterial lines. Our nursing protocols require tracheal suctioning every 4 hours and as needed for ventilated neonates, whereas a recent study⁵² showed comparable ventilatory outcomes when the frequency of routine suctioning was decreased to every 8 hours.

Nurses and physicians agreed that most neonatal procedures cause moderate or severe pain,⁴⁵ with pain scores above 4 estimated for 26 of the 31 procedures. Physicians ascribed lower pain scores; because they are mostly responsible for prescribing analgesics, this may contribute to the limited use of analgesic therapy in neonates.² Others have reported comparable significant differences in pain scores of invasive procedures between nurses and physicians.^{53 - 54} The clinicians not returning the questionnaire might have been unmotivated and not interested in pain management in the NICU. However, a formal analysis of their motives was not undertaken. This lack of interest is probably related to their belief that these daily procedures are not that painful. As the response rate was not very high, the results of our questionnaire might show some overestimation of clinicians' overall opinion about the pain level of these procedures.

In our NICU, nurses tried to cluster procedures during routine nursing care, after which they comforted patients by touch, pacifiers, or positioning (eg, swaddling), whereas other behavioral and environmental approaches were used irregularly. Similar to our results, a multicenter study⁵⁵ in France also showed minimal use of analgesics and a lack of standardization in the pharmacological regimens used in the NICU. Kahn et al⁵⁶ reported a 28.6-fold variation in the use of opioids among 6 NICUs. Variations in attitude toward pain may limit the generalizability of our study findings to other centers around the world.

Restrained use of opioids by neonatologists can be explained by the fact that there is wide disagreement as to whether the evidence base is sufficient to justify prolonged exposure to opioids in this vulnerable population. There is some evidence, from studies in rats, that neonatal morphine exposure causes specific long-term behavioral effects⁵⁷ and might cause retarded growth and motor development.⁵⁸ Underlying pathologic mechanisms have been demonstrated by morphine-induced apoptosis in human fetal cell cultures⁵⁹ and by µ-opiate receptor down-regulation following morphine treatment in neonatal rat brain.⁶⁰ Prolonged use of high doses of opioids in animal and in vitro models complicates extrapolation of these findings to daily NICU practice. The only study⁶¹ investigating long-term effects of human neonatal morphine treatment showed no effects in 5- to 6-year-olds. Anand and colleagues^{62 - 63} reported decreased mortality with the use of postoperative analgesia in premature infants. However, a similar benefit of routine use of morphine has not been reported in this population without a surgical operation. Grunau et al²¹ found that altered pain responses in preterm neonates were predicted by the number of previous painful procedures and were normalized by the early use of morphine analgesia. Although accumulating data suggest that analgesic therapy with morphine might be useful to prevent some of the long-term effects of repetitive neonatal pain, further evidence about the safety of prolonged use of opioids is needed.

We recommend that a continuous intravenous infusion of opioids should be considered for infants requiring respiratory support during the first 24 hours of admission, in combination with well validated pain scores as part of routine nursing care. Use of an algorithm would enable caregivers to respond immediately and in a structured way when they observe pain in these infants. Although this recommendation is supported by preliminary studies,^{39 ,64 - 67} results from larger randomized controlled trials are needed to decide if ventilated neonates should be routinely treated with continuous opioids.⁶⁸ As treatment regimens evolve, consistent practices in the NICU should be developed to minimize invasive procedures that continue to hurt our newborns.

Corresponding author: Dick Tibboel, MD, PhD, Department of Pediatric Surgery, Erasmus MC–Sophia Children's Hospital, Dr Molewaterplein 60, 3015 GJ Rotterdam, the Netherlands (e-mail: j.illsley@erasmusmc.nl).

Accepted for publication June 19, 2003.

This work was supported by grant MW-NWO 940-31-048 from the Netherlands Organization for Scientific Research, The Hague.

We thank N. Jongeneel, RN, C. Bunkers, RN, and E. Smink, RN, for distributing the surveys among physicians and nurses; C. Hoving, MSc, for helping to collect the data from nurses and physicians who completed the survey; H. J. Duivenvoorden, PhD, for statistical advice, and J. Hagoort for his help in preparing the manuscript.

As evidence about the short- and long-term harmful effects of pain in neonates accumulates, efforts should be made to decrease the number of pain experiences and to improve analgesic therapy. This study reports the number of painful procedures that neonates experience during an NICU stay. In contrast to the high incidence of painful procedures, we show the limited use of analgesic treatment, despite the awareness of nurses and physicians that most procedures are indeed painful.