Introduction

Central venous catheter placement consists of the in sertion of a biocompatible catheter into the central venous circulation for patient treatment and monito ring¹. These vascular accesses are an essential tool for the care of children in a pediatric intensive care unit (PICU), as they allow the administration of me dications, fluids, parenteral nutrition, or blood pro ducts, as well as laboratory sample collection and he modynamic monitoring^1,2. There are different classi fications of central venous catheters. According to the access to the central circulation, if the catheter is placed in a central vein (jugular or subclavian) it is called central venous catheter (CVC); if it is placed in the femoral vein, it is referred to as a deep femoral insertion venous catheter; however, if the catheter is placed in a peripheral vein, it is known as a peri pherally inserted central catheter (PICC)³.

CVC-related infection is the main cause of hospital acquired infection in pediatric ICUs. In a study conducted in the Pediatric Intensive Care Unit of the Centro Hospitalario Pereira Rossell in 2019, the incidence was 0.97 cases per 1,000 CVC days⁴. The diagnosis is initially based on clinical suspicion in the presence of local or general signs of infection, but these have low sensitivity and specificity. Therefore, microbiological techniques are required for the definitive diagnosis of catheter-related infection⁵. We refer to CVC-related sepsis based on IDSA criteria in a patient with an intravascular catheter with at least one positive blood culture for a known pathogenic microorganism. The sample is obtained through peripheral puncture, with clinical signs of systemic infection, with no other apparent source of infection, and with one of the following conditions:

If the catheter has been removed: the culture of the catheter tip and peripheral blood both being positive for the same microorganism (by quantitative o semiquantitative technique).

If the catheter has not been removed: a differential culture shows a higher bacterial load in the blood sample collected from the catheter lumen compared to the peripheral blood culture and both being positive for the same microorganism. The central culture is considered significantly higher than the peripheral one if the bacterial count of the detected pathogenis more than three times higher in the central blood culture or if the microorganism growth is detected (by automated blood culture system) at least 2 hours earlier in the central blood culture than in the peripheral one^5,6.

Infrared (IR) imaging at the bedside has great potential for use as a semiological clinical tool, since it is a non-invasive method that does not cause adverse effects, does not require sedation or moving the critically ill patient, and does not involve ionizing radiation^7-9. It allows a qualitative and quantitative analysis of the thermal radiation emitted, producing a high-resolution image called a thermogram^7-10. In the last 20 years, its use has increased exponentially, providing evidence of its contribution to various applications such as rheumatology, plastic surgery, palliative medicine, vascular pathology, diabetic foot, neoplasms, cardiology, ICU, occupational medicine, pain therapy, toxicology, and sports medicine^8-12. It has high sensitivity and specificity for providing information on metabolism, perfusion, and inflamma tion in a region of interest (ROI)^10-13. For accurate interpretation, a thorough understanding of the patho physiology, thermodynamics, and thermokinetics of human skin is essential^10-14.

Skin temperature regulation is a complex system that depends on the blood flow rate, local subcutaneous tissue structures, sympathetic nervous system activi ty, environmental factors, and the basal metabolic rate of the individual^15-17. Imaging in the ICU or the emergency room is performed at the patient's bedside with a defined protocol and scientific-grade thermal sensors to obtain thermograms. These are then analy zed using validated software for human thermo graphy by a physician trained in thermology. The objective of this study is to present two clinical cases of children with suspected CVC-related infection who were on follow-up using thermal imaging in a pediatric parenteral and enteral nutrition unit (UNEPP). The pathogenesis of the disease and clinical manifestations are analyzed, and infrared thermography is used for catheter monitoring.

Methodology

This report is part of a larger study that includes the follow-up of CVCs during PICU hospitaliza tion. It is an initial communication presenting two clinical cases included in the protocol, in which thermal images were recorded. The legal guardians of the patients consented to the reporting of these clinical cases, including their respective IR images. The protocol was approved by the Ethics Commi ttee of the Hospital Pediátrico Pereira Rossell and was registered with the Ministry of Public Health. A high-resolution FLIR E75 sensor (FLIR Systems AB, Taby, Sweden) was used, with a thermal sen sitivity of 0.03°C and a 307,200 pixels-resolution IR with UltraMax IR technology. Skin emissivity was set at 0.98. The patient underwent thermal acclimatization for 15 minutes before imaging. The room temperature was 24°C. The image acquisition protocol (UCINTERM) consistd of two thermal win dow approaches: one taken from 60 cm laterally and the other one from 60 cm frontally in relation to the patient unit, with the bed at a 45° angle, with safety strap, and with the patient seated. The 17-month-old patient was recorded in supine position. The 3-year and 2-month-old patient was also seated. The site of interest on the topography of the CVC was analyzed in the Pediatric Intensive Care Unit (PICU). Reference values were protocolized for the abdomen and contra lateral to the ROI (R1-R2). IR recordings were perfor med at seven-day intervals. Image capture and storage were performed with the VisionFy 1.2.1 platform (Thermofy). Once the thermal images were captured, ROIs were processed using the circle drawing tool, with quantification of temperatures through calculation of the delta T°C (∆T^max and ∆T^average) of the ROIs.

High-contrast rainbow color scales were used. 3D-IR reconstructions were processed to analyze ROIs and their thermal distribution. T°C delta value scales were defined according to the protocol of the International Consensus and Guidelines on Medi cal Thermography 2016-2018. Delta values 1°C severe dys function. International standards of the TISEM and Glamorgan protocols for thermal imaging were followed.

Clinical cases

Case 1: A 1-year and 5-month-old female with post-natal diagnosis of trisomy 21. The patient also presented with congenital heart disease (atrioven tricular canal with pulmonary atresia) for which a Blalock-Taussig shunt was performed at 5 days of life. She has hypothyroidism, currently treated with T4, and severe malnutrition with swallowing disor der, leading to admission to the UNEPP at CHPR with a mixed feeding regimen. The placement of a non tunneled left subclavian CVC was required to start parenteral feeding. Weekly thermographic and clini cal follow-up was conducted, showing a worsening of general condition, fever, and change in acute phase reactants. Blood and catheter culture were positive for multisensitive Staphylococcus epidermidis. It was decided to preserve venous access and initiate intravenous treatment with vancomycin along with a local vancomycin lock. The patient showed good clinical progress, with resolution of the infection and preservation of the CVC.

Case 2: A 3-year and 2-month-old preschooler with chronic intestinal pseudoobstruction syndrome lead ing to severe chronic malnutrition. Normal develop mental milestones for age. She was admitted to the UNEPP at CHPR on parenteral feeding. History of multiple hospital-acquired infections. Due to fever and change in acute phase reactants, empirical antibiotic therapy started with vancomycin and meropenem, with suspected CVC-related sepsis. Weekly ther mographic follow-up was performed. During illness, fever persisted, and multidrug-resistant Klebsiella pneumoniae was isolated from blood culture and ca theter culture. The CVC was re-moved, with subse quent clinical and thermal improvement.

Results

In case 1, the initial IR showed hyper-radiant areas in the selected ROIs along the path of the CVC (Figure 1 A, B, C). (Table 1 A, B). It shows ΔT^max values of 3.0°C and ΔT^average of 2.35°C in the ROIs. (Table 1 A) (Figure 2). The follow-up infrared imaging on day seven of the selected ROI showed a hyporadiant area along the path of the CVC compared to the initial assessment, with ΔT^max of 1.02 °C and ΔT^average of 1.06 °C in the ROIs, maintaining the patient's catheter. (Table 1 B). In case 2, the initial IR shows a hyper-radiant area along the path of the CVC, asymmetrical regarding the contralateral area in the ROIs. (Figure 3 A, B, C). The values obtained in the ROI showed a ΔT^max of 2.10°C and a ΔT^average of 2.06°C at the catheter inser tion site (Table 1 C). The catheter was ultimately re moved, and a new access was placed at a different site.

Figure 1 (A, B, C): Day 1. A: Thermal image at 45° position with safety strap. Arrows indicate ROI (r1) selected according to the UCIN-TERM protocol. B: Optical image. C: 3D thermal reconstruction, showing hyper-radiant areas. 

Figure 2 (A, B, C): Day 7. A: Thermal image in 45° position with safety strap. Arrows indicate ROI (r1) selected according to the UCIN-TERM protocol. B: Optical image. C: 3D reconstruction, showing hypo-radiant areas. Right lateral views at 60 cm. 

Figure 3 (A, B, C): Case 2: A: Thermal image in 90° sitting position. In the ROIs (r1), ΔT^max values of 2.08°C and ΔT^average of 2.04°C are observed at the catheter insertion point (arrows). B: 3D reconstruction, showing hyper-radiant areas (arrows). C: Optical image and ROI. Frontal views at 60 cm. 

Discussion

Hospital-acquired infections in the ICU lead to a marked increase in morbidity and mortality. Among hospital-acquired infections, CVC-related infections are the most prevalent and most often occur in critically ill patients⁵.

For the control of CVC-related infections, several factors must be considered: the type of catheter, signs related to catheter involvement, the clinical condition of the patient, and culture results⁶. When a patient with a CVC presents suggestive signs of bacteremia, such as fever, chills, and hypotension, the protocol in the ICU is to perform a catheter culture and a peri pheral blood culture at the same time. CVC-related infection is confirmed when the microorganisms from both cultures are identical in type and antibiogram, according to the criteria previously defined in the in troduction^1-6.

In cases of suspected CVC-related infection, some ICUs have protocols for its removal. However, there is evidence that catheters removed in these situations often do not grow pathogens in cultures. In addition, removing CVC in pediatric patients, especially critically ill ones, and even more so in children with prolonged parenteral feeding, can pose a healthcare challenge due to the difficulty of placing a new vascular access, and each new placement worsens their vital and functional prognosis. For this reason, it is necessary to have a methodology in the hospitaliza tion unit that can help in attempting conservative ma nagement of the CVC, at least initially, when dealing with an uncomplicated catheter-related infection^3-6. We say that a CVC-associated infection is compli cated when there is persistent bacteremia: growth of the same pathogen in blood cultures >72 hours after initiation of adequate antibiotic therapy, presence of endocarditis, septic embolisms, osteomyelitis, or sep tic thrombophlebitis^3,4. Internationally, in the last 20 years, multiple lines of research have been developed in the diagnosis and follow-up through thermogra phy, using protocols that allow for reproducible stu dies^7-16. It is used in the follow-up of multiple pathogies that present changes in perfusion, inflammation, and metabolism^11,13,17. IR has been used to study diseases where the skin temperature may reflect the presence of inflammation in underlying tissues or those in which blood flow increases or decreases due to a clinical abnormality(11-14). There are reports of the correlation between ΔT^average values in the ROIs and humoral infection markers^11,15,16. The described cases presented asymmetries and altered thermal dysfunctions with ΔT^average higher than 1°C^10-13. Additionally, both cases presented ΔT^max greater than 2.0°C^10,11,16. In a publication that performed thermal follow-up in a patient with a bothrops bite with local swelling, on the fourth day after the bite, an increase in the ΔT^average of up to 3°C was observed, concomitant with an increase in systemic leukocytosis and fever, with good response after starting antibiotic treatment, showing clinical, laboratory, and thermal improvement¹¹. Although we cannot extrapolate this finding to ΔT^average to other infectious processes, it could be indicative of processes related to systemic infections.

Another study performed a thermographic follow up in a hospital in Valencia on children with PICC with oncologic diseases. These children have the same difficulty in placing long-term venous access as the children in the clinical cases discussed with pa renteral feeding. In this group of patients, the tempe rature deltas were ΔT^max 1.5 ± 0.7°C in children with catheter-related sepsis, compared to ΔT^average 0.5 ± 0.2°C in children where no infection was confirmed. This difference was statistically significant, with a decrease observed once treatment was initiated¹⁶. The thermography team continues this project pros pectively with the follow-up of patients with CVC placement during their stay in intensive care, in order to increase the number of recruited patients and thus provide significant data for statistical conclusions, as well as define possible ranges of ΔT^average and ΔT^max that may be predictive of catheter-related sepsis in our setting^14-16.

Conclusions

IR thermography is a non-invasive study that can be useful for the follow-up and thermal semiological diagnosis of inflammatory processes in central catheters^4-6. There are few studies in the literature on ther mal follow-up of CVCs in pediatrics^14-16. This is the first study in Uruguay with high-resolution thermal imaging in pediatrics. CVC-related infections are the main cause of hospital-acquired infections in pedia tric ICUs. However, catheters removed due to suspec ted infection often did not show microbiological growth in cultures^4-6. The use of a technique that does not require mobilizing the critically ill patient, complementing the physical examination and humo ral parameters, together with monitoring ΔT^max/ ΔT^average values, combined with its low cost, shows promise. Further case series studies are needed for statistical analysis.