Chest CT in children: anesthesia and atelectasis
- First Online:
- Cite this article as:
- Newman, B., Krane, E.J., Gawande, R. et al. Pediatr Radiol (2014) 44: 164. doi:10.1007/s00247-013-2800-4
- 685 Downloads
There has been an increasing tendency for anesthesiologists to be responsible for providing sedation or anesthesia during chest CT imaging in young children. Anesthesia-related atelectasis noted on chest CT imaging has proven to be a common and troublesome problem, affecting image quality and diagnostic sensitivity.
To evaluate the safety and effectiveness of a standardized anesthesia, lung recruitment, controlled-ventilation technique developed at our institution to prevent atelectasis for chest CT imaging in young children.
Materials and methods
Fifty-six chest CT scans were obtained in 42 children using a research-based intubation, lung recruitment and controlled-ventilation CT scanning protocol. These studies were compared with 70 non-protocolized chest CT scans under anesthesia taken from 18 of the same children, who were tested at different times, without the specific lung recruitment and controlled-ventilation technique. Two radiology readers scored all inspiratory chest CT scans for overall CT quality and atelectasis. Detailed cardiorespiratory parameters were evaluated at baseline, and during recruitment and inspiratory imaging on 21 controlled-ventilation cases and 8 control cases.
Significant differences were noted between groups for both quality and atelectasis scores with optimal scoring demonstrated in the controlled-ventilation cases where 70% were rated very good to excellent quality scans compared with only 24% of non-protocol cases. There was no or minimal atelectasis in 48% of the controlled ventilation cases compared to 51% of non-protocol cases with segmental, multisegmental or lobar atelectasis present. No significant difference in cardiorespiratory parameters was found between controlled ventilation and other chest CT cases and no procedure-related adverse events occurred.
Controlled-ventilation infant CT scanning under general anesthesia, utilizing intubation and recruitment maneuvers followed by chest CT scans, appears to be a safe and effective method to obtain reliable and reproducible high-quality, motion-free chest CT images in children.
KeywordsChest CT Infants Atelectasis Lung recruitment Controlled ventilation
Chest computed tomography (CT) scans are invaluable in assessing the lungs in a wide variety of conditions in children. Modern CT technical advances have allowed for rapid scanning and decreased radiation dose while maintaining good anatomical definition for most routine CT scans of the chest . As CT scan speed has increased, the need for sedation for pediatric CT studies has dramatically decreased [2, 3, 4]. In spite of these advances, sedation or anesthesia is still required in many young children for optimal-quality CT in which detailed parenchymal evaluation is needed to detect early or subtle changes, or in which motion-free inspiratory or inspiratory plus expiratory chest CT scans are needed . This especially applies to infants/toddlers as well as uncooperative or intellectually or emotionally impaired children. There has been an increasing tendency for anesthesiologists to assume the responsibility of sedation/anesthesia for imaging . Individual anesthesia techniques vary widely and anesthesia-related atelectasis has proven to be a common and troublesome quality issue in our institution. Rescanning part of or the entire lung has been required in some cases with resultant concerns about unnecessary radiation exposure. The purpose of this study was to evaluate the reliability, safety and effectiveness of a standardized anesthesia lung recruitment technique that has been developed in our institution for CT chest imaging.
Material and methods
The study was approved by the institutional IRB and in each prospective case a signed parental informed consent was obtained.
General anesthesia was induced with inhaled sevoflurane using an anesthesia machine. A tight-fitting face mask was used during inhalation induction, adding continuous positive airway pressure as soon as consciousness was lost. Intravenous cannulation was then performed, and intravenous propofol and fentanyl were administered to augment the depth of anesthesia and permit peak inspiratory pressure (PIP) of 20 cm H2O and positive end expiratory pressure (PEEP) of 6 cm H2O. The trachea was intubated as soon as possible with a cuffed, appropriately sized endotracheal tube (ETT). Alveolar recruitment maneuvers were then completed, consisting of 10–12 3-s inspirations to a PIP of 40 cm H2O (32–35 cm H2O in first 6 scans) using a PEEP of 6 cm H2O between inspirations. This was followed by 3 scan-preparatory breaths at 25 cm H2O inspiration and PEEP of 6 cm H2O, followed by an inspiratory breath-hold at 25 cm H2O on the 4th breath for scan acquisition. The breath-hold was usually achieved by the spontaneous apneic pause produced after the previous hyperventilation and subsequent hypocarbia. This method was utilized for the scout and inspiratory CT scans. For the expiratory CT scan, a similar recruitment procedure was used except that upon exhalation after four preparatory breaths the patient was disconnected from the anesthesia breathing circuit (resulting in an airway pressure of 0 cm H2O) followed by a 4-s pause to allow for more lung deflation with acquisition of the expiratory scan near residual volume. Recruitment breaths were repeated in their entirety before the scout, inspiratory and expiratory scans (Fig. 1). When intravenous contrast was used for the inspiratory scan, the contrast injection was started at the end of the recruitment phase with initiation of the four 25/6 cm H2O scan-preparatory breaths.
All CT scans of the chest were performed on a multidetector CT scanner (Siemens Sensation 64; Siemens Medical Solutions, Malvern, PA). The CT scans were obtained with the following scan parameters: 80–100 kVp, 25–80 mAs, pitch 1.0–1.2. Inspiratory scans were spiral volumetric acquisitions (0.6-mm slice thickness), expiratory scans were acquired either in the spiral volumetric mode (0.6- to 1.2-mm slice thickness) or clustered axial high resolution CT gap mode (1-mm thick, 5- to 7-mm gap). All inspiratory and most expiratory scans could be acquired in a single breath-hold. If necessary, a second recruitment/breath-hold/scan sequence was done to complete the axial clustered expiratory images. The acquisition of very thin axial CT cuts was needed for the research study that used high-level postprocessing tools to evaluate the airways and lung; children with possible pulmonary metastases were recruited as control subjects. This is not the protocol used for either routine chest CT or HRCT at our institution.
Between March 2007 and July 2012, 56 controlled-ventilation CTs were prospectively obtained under anesthesia using this technique in 42 children: 24 girls, 18 boys; ages 2 months to 5.1 years, mean: 2.5 years. Seven children had multiple CT examinations using this technique, between 2 and 4 per patient. These 56 chest CTs constitute the protocol controlled-ventilation group for the current study. Five attending anesthesiologists with a wide range of experience were involved with these cases using this protocol.
Indications for chest CT included: cystic fibrosis in 11 children; primary ciliary dyskinesia in 4; chronic or interstitial lung disease in 17, and evaluate pulmonary metastases in 10. Intravenous contrast was utilized for the inspiratory scan in 11 examinations.
The radiology files of the same 42 children were retrospectively reviewed for additional chest CT scans performed under anesthesia during the same time period but without using this protocolized anesthesia/recruitment technique. A total of 70 chest CT scans in 18 children were identified constituting the non-protocol chest CT group. The lists of both protocol (56) and non-protocol (70) CT cases were combined and randomized on an Excel spreadsheet by a nonreader so that the scoring of the cases was done in a blinded fashion. Two pediatric radiologists, one faculty (B.N., 31 years of experience) and one pediatric radiology fellow (R.G.), independently reviewed and scored the reconstructed inspiratory axial lung algorithm images of all 126 cases on a 5 or 6-point scale for overall quality and atelectasis using the following ratings:
Quality: poor-1; satisfactory-2; good-3, very good-4 and excellent-5.
Atelectasis: none-0; minimal-1; small subsegmental-2; segmental-3; lobar or multisegmental-4, and whole lung-5. Low lung inflation was not considered atelectasis; there had to be actual pulmonary opacity, usually peripheral or dependent.
All scores from the two radiologists were averaged and were rounded up to the nearest integer.
The safety for these controlled-ventilation procedures was evaluated by clinical assessment during each procedure and recovery period, and by retrospective review of anesthesia records for each case. In 21 of the 56 controlled-ventilation CT cases, more extensive cardiopulmonary data were available for retrospective review due to the utilization of the comprehensive ventilatory parameters settings (SaAnesthesia, Version 2011.04.1.21; Cerner Corporation, Kansas City, MO, USA) during these cases, occasioned by the hospital conversion to a fully digitized anesthesia record with automated physiological data collection midway through the time period of this investigation. These data were compared with 8 conventional care control cases matched for age with the similar more extensive cardiopulmonary parameter evaluations. Seven of these 8 children were being evaluated for pulmonary metastases and one for recurrent pneumonia. Three of the 8 control cases were the same children who participated in a subsequent controlled-ventilation study (contributed to the 21 cases). Ventilatory data were not available for the expiratory chest CT scans for any of the cases, as these children were disconnected from the anesthesia monitoring circuit at the time of expiratory scanning.
Overall quality rating in protocol and non-protocol chest CT cases
Atelectasis rating in protocol and non-protocol chest CT cases
Comparisons of cardiorespiratory parameters between the 21 controlled-ventilation cases and the 8 matched conventional care cases were made with unpaired Student’s t-tests. Comparisons of cardiorespiratory parameters at rest, during recruitment maneuvers, and CT scan acquisition at 25 cm H2O in the controlled-ventilation cases were made using repeated-measure one-way ANOVA. Subsequent multiple pair-wise comparisons were completed utilizing the Tukey test. Statistical significance was determined at P<0.05.
Prospective study: 56 controlled-ventilation CTs in 42 children utilizing the standardized protocol outlined above. All 56 examinations were completed safely with no procedural complications; one child had an episode of emergence delirium during recovery that resolved without specific intervention. Since the scoring, as well as detailed cardiorespiratory evaluation, was obtained on the inspiratory CT studies, the expiratory imaging is not included in the results or statistics. A few examples of expiratory imaging are shown in the figures for illustrative purposes.
For agreement between readers, the Fleiss and Cohen weighted kappa (FCWK) coefficient for quality scoring was 0.79 with a 95% confidence interval of (0.73, 0.85) on a sample size of 126 CT cases. For atelectasis scoring, the FCWK coefficient was 0.92 with a 95% confidence interval of (0.88, 0.97) on a sample size of 126 cases. These findings demonstrated excellent agreement between readers, especially for the atelectasis score.
For overall comparison between protocol and non-protocol cases for quality score, the Rao-Scott chi-square test demonstrated a significant difference between the two (P<0.0001). For overall comparison between protocol and non-protocol cases for atelectasis score, the Rao-Scott chi-square test also demonstrated a significant difference (P=0.0175).
Mean (SD) of cardiorespiratory parameters in the protocol vs. control groups
In spite of a significant trend for lower respiratory rate and heart rate (ANOVA), in the group of 21 controlled-ventilation children when viewed serially at recruitment and inspiratory scan acquisition compared to baseline (Table 3), there were no significant differences noted in mean arterial pressure or oxygen saturation during these transitions compared to baseline. These data indicate that there were stable blood pressure responses and no significant cardiovascular compromise during recruitment and inspiratory breath-holds despite the short-term increased inspiratory pressures, hyperventilation and hypocapnia during these maneuvers. In these 21 controlled-ventilation cases, effective recruitment maneuvers were maintained at peak inspiratory pressure (PIP) values near 40 cm H2O (39 [2.0]). This translated to a mean (SD) tidal volume (Vt) of 487 ml (232.9 ml) or 35 ml/kg (14.0 ml/kg) (normal resting value 12 ml/kg) and a mean (SD) end tidal CO2 value of 24 mmHg (6.5 mmHg) (normal 30–45 mmHg). During inspiratory scan acquisition, the mean (SD) PIP value was 25 cm H2O (1.2 cm H2O). This corresponded to a lower mean (SD) Vt of 300 ml (166.0 ml) or 21 ml/kg (10.4 ml/kg), and a mean (SD) end tidal CO2 of 27 mmHg (8.7 mmHg).
A great deal of effort has been expended on developing techniques to obtain high-quality CT imaging of the chest without requiring a breath hold or patient sedation/anesthesia. These have involved the use of important resources such as child life personnel and included careful patient and parent education and preparation, the use of CT models, parental presence in the scanner, distraction techniques such as music, light and decorations as well as feed, wrap and scan protocols . Decreased need for CT sedation has resulted in decreased scheduling complexity, cost and morbidity related to sedation or anesthesia.
Nonetheless, there are data that confirm that motion-free inspiratory studies are optimal in children when detail of lung parenchymal anatomy is required, and early or subtle findings are being sought . Examples include early changes of cystic fibrosis lung disease or interstitial lung disease in infants and in infants and toddlers with small pulmonary metastases. Important errors of interpretation can be made when the quality of the scan is suboptimal and these may greatly affect patient management. In addition, when inspiratory and expiratory HRCT scans are done, it is important that they be obtained at the appropriate time in the respiratory cycle  and highly desirable for repeat studies to be obtained at known and reproducible inspiratory and expiratory lung volumes.
All pediatric CT sedation in our institution is provided by physician anesthesiologists and no additional patient financial charges were incurred by applying the controlled-ventilation technique. Anesthesiologist control of CT sedation is probably optimal for patient care, as anesthesiologists are better equipped to deal with the pharmacology of hypnosis and analgesia, control of the child’s airway, changes in physiology and emergent situations, allowing radiologists to concentrate on performing and interpreting imaging. However, it has been reported previously that the incidence of atelectasis is greater with anesthesia-monitored sedation (6.6%) than radiologist-monitored sedation (0.01%) [11, 12], and is consistent with the well-described atelectasis that occurs promptly with the administration of anesthetics . That has also been the subjective impression at our institution. This might be related to a deeper level of CNS depression, respiratory muscle relaxation, shallower breathing and ventilation/perfusion mismatch with positive pressure ventilation vs. light or moderate sedation.
There has been a great deal of discussion in the anesthesiology literature as to the nature of anesthesia-induced atelectasis and how best to prevent lung collapse and reinflate lung that is atelectatic. The prevailing theory of causation of anesthesia-induced atelectasis is loss of respiratory muscle function with reduced functional residual lung capacity [11, 14]. The effects of general anesthesia on atelectasis are far more apparent in infants and young children than in older children and adults, since functional residual capacity in infants is primarily maintained by active static and cyclic contraction of thoracic inspiratory muscles to maintain a much more compliant chest wall . Methods that have proven effective in preventing atelectasis include abstention from pre-anesthesia hyperoxygenation, which results in airway closure with resorption atelectasis in dependent regions of the lung , and use of PEEP (5–10 cm H2O) [16, 17]. Recruitment of atelectatic lung appears to be most effective with lung inflation to near vital capacity at 40 cm H2O, and this was much more effective than volume inflation including double tidal volume (sigh breaths) . The rationale for 40 cm H20 is that this is the alveolar opening pressure in healthy adults; this number may be even greater in children who have smaller airways and alveoli .
The intubation lung recruitment technique that we have developed, modeled on that of the AREST CF group, can be performed safely and consistently by different anesthesiologists using a standardized protocol. Procedure-induced atelectasis that affects quality is for the most part reliably absent and repeat sequences are not needed. It is important that the protocol be followed carefully. Key features are continuous PEEP (6 cm H2O) at the initiation of anesthesia, prompt intubation with a cuffed endotracheal tube and effective repeated recruitment efforts at relatively high PIP (38–40 cm H2O). Initially, because of concerns regarding the relatively high inspiratory airway pressures in young children and the risk of barotrauma, we used lower PIP (32–35 cm H2O) in the first 6 cases. Results were not optimal and it was decided to revert thereafter to the higher PIP (38–40 cm H2O) recommended by the Perth Australian group (AREST CF) .
There may be a few situations in which the PIP we used could be considered high risk, such as in a child with severe bronchopulmonary dysplasia or cystic lung disease. This situation would necessitate a discussion between the anesthetist and radiologist about modifying the protocol. There were some children in our group with chronic lung disease, and though our experience remains limited, we have not seen parenchymal lung injury, barotrauma or air leak either on CT scan or clinically, despite mean peak inspiratory pressures of 39 cm H2O, and tidal volumes approaching 35 ml/kg during recruitment maneuvers. During baseline, recruitment maneuvers, and inspiratory and expiratory CT scan acquisitions, there was no hemodynamic instability. During recruitment maneuvers and inspiratory scan acquisitions, there was relative hypocarbia secondary to transient hyperventilation, but this was of short duration, did not lead to adverse events or outcomes and facilitated apnea during image acquisition. In young children being evaluated for lung disease, our limited sample suggests that controlled-ventilation infant CT scanning is a useful diagnostic test with a reasonable and reassuring safety profile. This has also been confirmed with the large number of studies that have been completed in young CF children by the AREST CF group in Australia (6).
We did not specifically evaluate peri- and post-anesthesia clinical outcome measures, but no adverse events were observed or reported in this small sample of 56 scans in 42 children followed closely in pulmonary and oncology clinics within the institution. Therefore, we can state that the overall risk is no worse than approximately 5% based upon this sample size, , and is likely much lower. Indeed, the inherent risks of anesthesia are very small but do include drug reactions (both allergic and idiosyncratic), hypotension in fasted/dehydrated children, hypoxia with loss of airway control, laryngeal or tracheal trauma and edema from intubation, emergence delirium, gastric aspiration with or without subsequent pneumonitis, and secondary brain damage or death. In addition, there is a speculative risk of an adverse effect of some or all general anesthetics on neuronal development in infants and very young children, a concern that has been shown to occur under laboratory conditions but awaits confirmation of its existence in clinical medicine. While intuitively, moderate sedation seems less invasive and safer than general anesthesia, it shares the same risks enumerated above with the exception of laryngeal/tracheal trauma but carries a higher overall risk of hypoxia/loss of airway control with attendant secondary effects, as well as gastric aspiration, pneumonia and death. Although unintended, moderate sedation not infrequently becomes deeper to the level of general anesthesia with still greater attendant risks in the less experienced hands of the non-anesthesiologist. These risks need to be balanced against the risks of poor scan quality and additional radiation exposure if repeat imaging is needed, and the substantial risk of missed or misinterpreted findings.
The retrospective review of non-protocol cases clearly shows that when individual methods were used, even by experienced anesthesiologists with or without endotracheal intubation and aware of the need to prevent atelectasis, the results were unreliable, resulting in 75% of these cases having subsegmental, segmental or multisegmental/lobar atelectasis and repeat studies being needed in 5 cases. Note that the decision to repeat a non-protocol study was not decided by these authors but was left to the supervising radiologist of the study with no specific criteria for deciding how much atelectasis rendered the examination uninterpretable. The decision not to repeat any of the protocol cases was made by the study radiologist present at the scan, again without specific criteria. However, since this was a research study, it is likely that the standards for deciding that an examination was adequate were quite high.
Airway management using a laryngeal mask airway is a common anesthesia technique. However, this method does not permit sufficiently high lung recruitment pressures and, in our experience, was unreliable with the additional undesirable result of gaseous gastric distention (Fig. 4). Intubation alone or with the addition of PEEP but without recruitment was also insufficient to prevent atelectasis.
One relative weakness of the current study is that the 70 control CT studies were not chosen randomly, which could create some bias since those children who had previously had poor-quality CT might have been recruited subsequently into the controlled-ventilation group. Viewed from another perspective, the success of the recruitment controlled-ventilation protocol in such cases further indicates how useful it was to reliably produce high-quality scans. A similar but opposite bias may have occurred in that the anesthesia technique was chosen by the anesthesiologist in the control group and they could have used a similar recruitment technique to the controlled-ventilation protocol. As our study progressed, we found increasing anesthesiologist comfort with the protocol and other anesthesiologists began to adopt the method as an effective way of managing children outside of the research study. With practice, the technique has become easier and faster. Current CT table time is approximately 15 min including induction of anesthesia, intubation, recruitment and inspiratory/expiratory scanning, emergence from anesthesia and tracheal extubation.
Other investigators have described controlled-ventilation techniques utilizing intravenous or oral moderate sedation and face-mask hyperventilation to obtain the reflex apneic pause at relative inspiratory and expiratory volumes . We have found this method to be much less reliable with poor reproducibility of inspiratory and expiratory volumes, less consistency in the absence of atelectasis, elevation of the diaphragm because of gastric distention and a risk of gastric aspiration in spite of attempted cricoid pressure by the anesthesiologist.
Our research technique can readily be adapted to routine helical chest CT or HRCT in infants and young children who require sedation/anesthesia. While it is always our goal to obtain optimal scans in all patients requiring chest CT, we are currently undertaking a focused quality improvement initiative in our department for pediatric chest CT under anesthesia. We plan to implement the intubation recruitment protocol for indications where good lung inflation is particularly important and atelectasis can really confound diagnosis; these include evaluation for pulmonary nodules or metastases, interstitial lung disease, congenital lung abnormality, pneumonia and conditions associated with air trapping or bronchiectasis.
Controlled-ventilation CT scanning utilizing a cuffed ET tube and airway recruitment maneuvers followed by inspiratory and expiratory CT scans under general anesthesia is a safe and feasible method to obtain reliable and reproducible high-quality, motion-free chest CT images in young children.
Source of funding: The project was supported by the National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health, through UL1 RR025744, and by the Lucile Packard Foundation for Children's Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Conflicts of interest