Pain-related changes in crSO2 among premature infants undergoing PICC insertion

Abstract Objective To evaluate the effectiveness of cerebral regional oxygen saturation (crSO2) values, measured using near-infrared spectroscopy (NIRS), in assessing pain associated with the peripherally inserted central catheter (PICC) in premature infants. Methods NIRS was used to measure the crSO2 levels of 48 premature infants with gestational age (GA) of < 32 weeks or a birth weight of < 1500 g. Premature infant pain profile (PIPP) scores, vital signs, transcutaneous oxygen tension (TcpO2), transcutaneous carbon dioxide tension (TcpCO2), and crSO2 values were monitored. One-way repeated measure analysis of variance was used to compare heart rate (HR), respiratory rate (RR), blood pressure (BP), peripheral oxygen saturation (SpO2), TcpO2, TcpCO2, and crSO2 values before (Time 1), during (Time 2), and after (Time 3) PICC insertion. The correlation between the PIPP scores at Time 2 and the fluctuations (values detected at Time 2 minus those at Time 1) of SpO2, TcpO2, and crSO2 were also analyzed. Results The PIPP score at Time 2 was significantly higher than those at Times 1 and 3. HR, RR, and BP values increased (p < .05), and SpO2 and crSO2 levels decreased at Time 2 (p < .05) compared with those at Time 1. Stratified analysis based on GA revealed significant differences in HR, RR, and crSO2 values between Times 1 and 2 in infants with a GA of ≥ 32 weeks. In infants with a GA < 32 weeks, significant differences were observed in HR, RR, SpO2, BP, and crSO2 values between Times 1 and 2. The fluctuation of the crSO2 level was strongly correlated with the PIPP score at Time 2 (r = −0.829, p < .001). A weak correlation was observed between the PIPP score at Time 2 and TcpO2 level fluctuation (r = 0.375, p = .009). No correlation was observed between the PIPP score at Time 2 and SpO2 level fluctuation (r = 0.242, p = .097). Conclusion The fluctuation of crSO2 levels strongly correlates with PICC procedural pain. Hence, crSO2 levels measured using NIRS may be used as an indicator for pain assessment in premature infants.


Introduction
Advances in medical technology and care have improved the survival rates of most preterm infants with gestational age (GA) between 24 and 32 weeks [1].However, this is accompanied by an increase in pain-producing invasive treatment procedures [1,2].A study analyzing the frequency of invasive procedures in the neonatal intensive care unit (NICU) discovered that newborns undergo 7.5 to 17.3 procedures daily during their hospital stay, with the most common procedures being heel lancing, suctioning, venipuncture, and peripheral venous catheterization [3].The peripherally inserted central catheter (PICC) may be placed and retained over a long period, increasing the safety of some medication infusions, and helping avoid frequent punctures; it is commonly used in the NICU, in particular, in the treatment of very preterm and very low birth weight infants who require high-osmolarity parenteral nutrition or intravenous access for longterm [4][5][6].As an invasive procedure, PICC insertion inevitably causes pain in infants.A study reported that the premature infant pain profile (PIPP) score of preterm infants during PICC insertion reached 11.74, representing moderate pain [7].
The nociceptive (pain) system starts developing during fetal life.Peripheral cutaneous sensory receptors begin appearing at 7 weeks of GA, and pain neurotransmitters are secreted after 8 weeks of GA; the first brain fold appears at 13-15 weeks of GA and later develops into the insula-one of the most important areas for pain perception; pain emerges at approximately 20-22 weeks of GA as a neuroadaptive phenomenon [8,9].Premature infants are highly sensitive to pain [10], and pain exposure in early life has been associated with short-and long-term adverse outcomes, including acute behavioral, metabolic, endocrine reactions, neurodevelopment, and regulation of stress systems.Furthermore, accumulating evidence shows that repeated invasive procedures cause central sensitization and hyperalgesia, leading to long-term changes in pain processing and the child's development [1,11,12].Moreover, the wind-up phenomenon proposes that repeated harmful stimuli over time can amplify detrimental effects on neonatal brain development, allowing harmless stimuli to be perceived as painful [13].
Accurate pain assessment can better prevent and treat the pain of preterm infants.Presently, various scales are commonly used for bedside pain assessment of newborns.These scales are based on changes in behavioral performance and physiological responses of newborns, such as facial expression, body movement, heart rate (HR), and oxygen saturation (SpO 2 ) [14].Owing to the immature regulatory system at each central nervous system level in premature infants, the behavioral and physiological responses to pain stimulation can be minimal and less noticeable, especially in infants with small GA [15][16][17][18].Moreover, infants' unresponsiveness to painful stimuli makes pain scale assessment challenging, as many variables, such as diagnosis or disease severity, can influence it [1].The invention of Near-infrared spectroscopy (NIRS) solves these problems and has become an important tool for assessing neonatal pain by measuring the cortical hemodynamic response to pain stimuli [19].Brain activation caused by pain stimuli increases oxygen consumption [20], and cerebral blood flow is highly correlated with facial expressions of pain [21].NIRS offers noninvasive, real-time monitoring of cerebral blood flow volume and cerebral oxygenation [22].Therefore, given the limited ability of premature infants to self-regulate cerebral blood flow and their vulnerability to brain injury, NIRS can serve as an adjunct in pain assessment.Additionally, the clinician's consideration of the preterm infant's medical history and current physiological status provides a new dimension for a more detailed and accurate assessment of pain in premature infants [22].
However, few previous studies have examined the application of cerebral regional oxygen saturation (crSO 2 ) levels measured using NIRS, in evaluating pain associated with PICC use in premature infants.Therefore, this study aimed to explore the correlation between crSO 2 levels and PIPP scores in premature infants undergoing PICC to provide a new indicator for pain assessment in the NICU.

Patient selection
This study was conducted at the NICU of the Affiliated Hospital of Southwest Medical University, Luzhou, Sichuan, China, from January to October 2021.Inclusion criteria involved preterm infants with GA < 32 weeks or a birth weight < 1500 g and those who underwent PICC insertion within 3 days after birth.
Exclusion criteria included infants with one or a combination of the following: severe brain injuries caused by asphyxia, birth injury, intrauterine infection, or other factors; congenital metabolic diseases; severe congenital heart disease, nervous system malformation, or other congenital diseases or severe complications; and imprecise or unrecordable data due to crying, struggling, or nonstandard use of instruments.
The ethical committee of the Affiliated Hospital of Southwest Medical University approved the study (ref approval no.KY2019248), and informed consent was obtained from all infants' parents.

PICC placement
The PICC (1.9 French; Becton, Dickinson and Company, REF368100, Paramus, New Jersey) was inserted following the guidelines recommended by the Infusion Nurses Society [23,24].The puncture sites were selected based on clinical need and vein accessibility in the infants, with a preference for lower limb insertion when the patient's condition permitted it.The length of the body surface was measured from the puncture point along the groin to the umbilicus and then to the midpoint of the xiphoid for lower limb insertion.The measured length was from the puncture site to the right sternoclavicular joint and down to the second intercostal space for upper extremity insertion [25].Certified doctors and experienced nurses set up a PICC workstation at the infant's bedside under sterile conditions, minimizing the number of needle punctures to improve success rates.The skin around the puncture site was disinfected before catheterization, and the position of the catheter tip was confirmed by X-ray after catheterization.Then the transparent dressings were applied around the puncture site for fixation and infection prevention.The procedure was successful when the catheter tip was placed in the inferior vena cava above the diaphragm (about the eighth to tenth thoracic vertebra) or at the lower third of the superior vena cava (about the fourth to fifth thoracic vertebra) [25].Finally, nurses recorded the type of catheter used, insertion time, insertion site, insertion length and vessel, catheter tip position, and any complications during the insertion.The nurse did not simultaneously administer other pharmacological or non-pharmacological pain relief measures to accurately assess pain in preterm infants during PICC intubation.

Observation time point
Data collection was performed at three time points.Before insertion (Time 1): 15 min before the insertion started when the infant was in deep sleep; during needle insertion (Time 2): from the puncture needle breaking the skin to the needle placement; and after insertion (Time 3): 15 min after the procedure when the infant was returned to a comfortable position and had recovered normally (infants stopped crying, gradually fell asleep, and HR and SpO 2 ceased fluctuating widely).

Outcomes and measurements
Primary outcomes 1. Regional cerebral oxygen saturation (crSO 2 ) level was measured using NIRS (EGOS-600A/B/C nearinfrared blood oxygen monitoring instrument, Suzhou FNGINMED, Beijing, China).NIRS is a measurement based on optical principles, with a 3 kHz sampling rate and near-infrared light at 700-1000 nm wavelengths, which can penetrate through the thin skin, scalp, and skull of infants [26].Hemodynamic fluctuations caused by neuronal activity can be monitored by positioning NIRS across relevant cortical volumes [27].The probe consists of a light source and two sensors, with a distance of no more than 2-3 cm between them being sufficient for newborn infants [28,29].
The nurse observed the forehead skin of the infant and cleaned it before the measurement.Subsequently, the nurse gently fixed the instrument probe in the middle of the infant's forehead with an elastic bandage and recorded the measurements.The probe was promptly removed after the measurement, and the infant's skin condition was observed.The skin was immediately treated if damaged.2. Premature Infant Pain Profile (PIPP) was used to assess pain in premature infants.This scale had good construct validity and moderate internal consistency [30,31].The maximum total score is 21 points, with higher scores indicating more intense pain.Nurses calculated the PIPP scores by observing the infants' facial expressions and behavioral status during each period.
Secondary outcomes 1. Heart rate (HR), respiratory rate (RR), blood pressure (BP), and peripheral oxygen saturation (SpO 2 ).These vital signs were continuously measured with a Mindray BeneView T5 Monitor (Mindray Bio-Medical Electronics Co., LTD, Keji 12th Road South, High-tech Industrial Park, Shenzhen, China).2. Transcutaneous oxygen tension (TcpO 2 ) and transcutaneous carbon dioxide tension (TcpCO 2 ) values were monitored using TCM4 (Radiometer Medical AsP, Denmark), an instrument that avoids repeated punctures for blood collection.The nurse attached a fixing ring to the infant's chest, dropped contact fluid, fixed the electrode, and recorded the measurements.These outcomes, except the PIPP score, were recorded after the monitor had worked steadily.Finally, the averages of the three stabilized values for Times 1, 2, and 3 were calculated to be used in the analysis.

Nurse training
The doctors with more than 5 years of clinical experience were selected, nurses with PICC puncture qualifications, and other nurses specializing in indicator monitoring to form a research team before the study and conducted four teaching training sessions.The training included the study design, the operation of NIRS and TCM4, and the use of the PIPP scale.Theoretical classes were taught through videos, PowerPoint presentations, and group discussions.Practical classes were taught using live demonstrations and one-by-one exercises.Finally, an examination was conducted to select the most suitable researchers for this study.

Statistical methods
Statistical analysis was performed using SPSS 19.0 (SPSS Inc., Chicago, IL).Categorical data were described as percentages (%).Based on their data distribution, continuous variables were reported as means ± standard deviations (X ± SD) or medians and interquartile ranges (P 50 [P 25 , P 75 ]).One-way repeated measure analysis of variance was used to compare HR, BP, RR, SpO 2 , TcpO 2 , TcpCO 2 , and crSO 2 values before, during, and after the insertion.The Bonferroni post-hoc test was used for pairwise comparisons between the groups.Pearson correlation analysis was used to explore the correlations between the PIPP score during PICC insertion and SpO 2 , TcpO 2 , and crSO 2 level fluctuations (values detected at Time 2 minus those at Time 1).

Results
During the study period, 75 premature infants had a GA of < 32 weeks or a birth weight of < 1500 g.Among them, 66 infants underwent PICC insertion; however, 13 were excluded for undergoing the procedure after 72 h of life.Additionally, 3 infants were excluded because their caregivers declined participation, and 2 infants with congenital heart disease were excluded.Finally, 48 infants were included in this study.The data of 48 infants were examined before statistical analysis and found no imprecise data.Therefore, all data were statistically analyzed.
The study population had a mean GA of 31.4 ± 2.2 weeks and a mean birth weight of 1294 ± 163 g.Apgar scores at 1 min and 5 min were 7.5 ± 2.2 and 8.5 ± 1.6 points, respectively.Among all infants, 95.8% received noninvasive ventilation support.The mean age at PICC insertion was 52.6 ± 16.7h.The mean weight of participants who underwent PICC insertion was 1208 ± 163 g (Table 1).

PIPP scores before, during, and after PICC insertion
The PIPP scores during PICC insertion significantly increased in all premature infants (p < .001).No significant difference was observed in the PIPP scores between preterm infants with a GA of !32 weeks and those with a GA of < 32 weeks at any time point (Table 2).

Physiological parameters before, during, and after PICC insertion
Figure 1 shows that significant fluctuations in HR, RR, SpO 2 , BP, and crSO 2 values were observed in all infants group and GA < 32 weeks group during PICC insertion.Premature infants with a GA of !32 weeks had significant fluctuations in HR, RR, and crSO 2 levels during PICC insertion.

Discussion
Pain is often described in a self-reported manner, but this is not possible for infants.Scales based on measurable behavioral and physiological indicators are used to assess pain in infants [32].In this study, the PIPP scores ranged from 10 to 14 points during PICC insertion, reflecting moderate to severe pain, consistent with the finding of a previous study [7].Acute pain is usually accompanied by significant fluctuations in physiological indicators, including HR, RR, BP, and SpO 2 , as discovered in this study.However, physiological changes are not specific to pain because many other external and internal factors, such as age, noise, bright light, mechanical ventilation, hypovolemia, or fever, can also contribute to these changes [33].Therefore, assessing pain using vital signs alone is insufficient.The perception of pain requires higher levels of cortical processing, the combination of behavioral physiological indicators and cortical activity can help determine pain in preterm infants, such as NIRS [22,34].A study using NIRS to measure hemodynamic activity in infants aged 25-43 postmenstrual age (PMA) discovered that the PIPP score correlated   ÃÃÃ P < .001 between time 1 and time 2 in the three groups, respectively; ### P < .001 between time 2 and time 3 in the three groups, respectively; Ã P < .05 between time 1 and time 2 in all infants group and gestational age < 32 weeks group.
well with cortical activity, and cortical pain responses were detected in some infants with no facial expression changes [21].Another study used NIRS to measure changes in total hemoglobin concentration (HbT) in infants aged 25-45 PMA when subjected to noxious and non-noxious stimulation.It reported that the mean contralateral HbT increased by 7.74 mmol/L after the harmful stimulus, and early cortical response after pain stimulation measured using NIRS may be more sensitive and specific than spinal-mediated reflex limb changes [35].In this study, NIRS was used to measure changes in crSO 2 during PICC insertion in preterm infants, discovering that crSO 2 decreased significantly after PICC insertion, with a strong correlation observed between crSO 2 level fluctuations and PIPP scores during PICC insertion.In addition, we compared the fluctuations of some oxygenation-related indicators, including SpO 2 and TcPO 2 , and observed a weak correlation between the PIPP score during insertion and TcpO 2 level fluctuation.However, no correlation was observed between the PIPP score and SpO 2 level fluctuation.This evidence suggests that crSO 2 is a more sensitive indicator of pain levels than SpO 2 and TcPO 2 .These findings further confirm the conclusion of the above study that the cortical response to pain in the early stage is more sensitive than the spinal cordmediated pain response.
Ercan et al. used the NIRS to measure the crSO 2 of preterm infants before and after PICC insertion and discovered that 46% of infants showed a significant decrease in mean crSO 2 within the first 15 min after PICC insertion.Over the next 15 min, only 22% of infants experience a significant decrease in mean crSO 2 compared with baseline crSO 2 levels [36].This suggests that PICC insertion may negatively affect cerebral perfusion, leading to a transient decrease in crSO 2 .In addition, the study discovered that the median time for crSO 2 levels to return to baseline after PICC insertion differed based on GA.Infants < 28 weeks required 25 (15-60) min for their crSO 2 levels to return to baseline, compared with 17.5 (3.7-35.7)min for those !28 weeks [36].In this study, the time required for crSO 2 levels to return to baseline in preterm infants of different GAs was not compared.However, as shown in Figure 1, the crSO 2 of the small GA group was lower at time 3, meaning that small GA infants needed a longer time to recover.Moreover, this study also found that physiological parameters fluctuate more in premature infants with GA < 32 weeks, especially the SpO 2 and BP, after stratification using GA.This may suggest that pain from PICC insertion had a greater effect on smaller premature infants.Premature infants with a lower GA may have poor cerebral autoregulation and are less likely to adapt to pain quickly.
There may be two reasons for the low crSO 2 in preterm infants.One is the decreased oxygen delivery to the brain due to hypocarbia, patent ductus arteriosus (PDA), severe anemia, hypotension, and desaturation (SpO 2 < 85%).Another is the increased brain oxygen consumption due to pain, convulsion, and fever [36,37].The decrease in cerebral oxygen in the study may be related to the increase in cortical oxygen consumption.According to the combination of previous studies and this study, crSO 2 is a sensitive indicator of physiological changes after pain stimulation.However, some limitations are remaining in routine clinical pain monitoring.First, NIRS is not universal and has not been introduced in some smaller hospitals, and nurses must undergo professional training.Second, NIRS is easily affected by infant activities, resulting in inaccurate readings, so multiple readings are needed to obtain the average to reduce errors.Therefore, compared with changes in behavioral and physiological indicators that are easier to obtain and observe, the practicability and acceptance of NIRS to measure cortical pain response still need to be further studied.
This study has some limitations.First, it was an observational study, which may have been affected by selection bias.Second, this study included infants intubated within 3 days of admission, finally only included 48 infants, which may have affected the reliability of findings because this was a small sample size study in a single center.Third, since the PIPP-Revised (PIPP-R) scale has not been translated into Chinese nor verified its content and construct validity in our hospital, the PIPP scale was chosen instead of the PIPP-R scale, which was also a limitation of this study.
Preterm infants experienced moderate pain during PICC insertion, physiological indicators fluctuated, and crSO 2 levels significantly decreased.A strong correlation was observed between crSO 2 level fluctuations and PIPP scores.Therefore, pain response at the cortical level can sensitively indicate pain levels in preterm infants.NIRS can be used as a potential pain assessment tool to better assess pain in preterm infants when combined with the PIPP scale.In future studies, the relationship between the degree of crSO 2 change and pain intensity, as well as the changes in the body's response to pain in preterm infants when crSO 2 gradually returns to normal will be explored.

Table 1 .
Baseline characteristics of the study participants.GA: gestational age; FiO 2 : fraction of inspired oxygen; PDA: patent ductus arteriosus.

Table 2 .
PIPP scores before, during, and after PICC insertion.