Effect of Carbon Dioxide Pneumoperitoneum on Development of Atelectasis during Anesthesia, Examined by Spiral Computed Tomography

After approval by the Ethics Committee of the Karolinska Institute, Stockholm, Sweden, seven patients scheduled to undergo laparoscopic cholecystectomy were included in the study after written informed consent was obtained. There were six women and one man, with a mean age of 40 yr (range, 34–51 yr), a mean body weight of 80 kg (range, 55–99 kg), and a mean height of 170 cm (range, 157–177 cm). All patients but one had a normal body mass index (mean, 27.6 kg/m²). One patient had a body mass index of 33.9 kg/m². All patients lacked history or signs of cardiopulmonary disease (American Society of Anesthesiologists physical status I). Preoperative spirometry and chest radiography did not reveal any signs of respiratory disease. The patients fasted from midnight the night before surgery. The operations were started in the morning. The study was prescheduled for 10 patients. Because all seven patients responded similarly after the induction of pneumoperitoneum, the number of patients was reduced for ethical reasons.

Patients were premedicated with 5–10 mg intramuscular ketobemidone. After administration of 0.2 mg glycopyrrolate, anesthesia was induced with 2–3 mg/kg body weight propofol, 0.15–0.2 mg fentanyl, and 6– 10 mg cisatracurium. Anesthesia was maintained by continuous infusion of 8–10 mg · kg⁻¹body weight · h⁻¹propofol. The patients were mechanically ventilated with a mixture of oxygen and air with a fraction of inspired oxygen of 0.35–0.40 by a ventilator (Siemens Servo Ventilator 900C; Siemens-Elema, Solna, Sweden). The ambition was to not influence breathing mechanics by changing ventilator setting during the investigation. Therefore, the patients were ventilated to an end-tidal carbon dioxide in the lower range of the normal interval at the start of the investigation, which required a tidal volume of 8 ml/kg body weight. No ventilatory adjustments were then needed to keep end-tidal carbon dioxide within the normal range during the measurements. During the investigation, an isotonic crystalloid solution was given at a rate of 2 ml · kg⁻¹body weight · h⁻¹(Glucose Baxter, 25 mg/ml with 270 mOsm/l (Baxter Medical AB, Kista, Sweden).

At measurements, heart rate and noninvasive mean arterial pressure were recorded by a Siemens Sirecust monitoring device (model 1281; Siemens-Elema). The fractions of inspired oxygen and carbon dioxide and the end-tidal fraction of carbon dioxide, as well as the percutaneous oxygen saturation, were monitored by a Datex Capnomac Ultima monitoring device (ULT-S-23-01; Datex Instrumentarium Corp., Helsinki, Finland). The respiratory rate and peak airway pressure were recorded from the ventilator.

The first measurement and CT scan were made after 27 ± 4 min of stable anesthesia with the patient in the supine position, with the arms lifted cranially on the tomograph table. Thereafter, sterile conditions were obtained, the laparoscope was introduced, and carbon dioxide was insufflated into the peritoneal cavity until the intraabdominal pressure reached a level of 11–13 mmHg, which was maintained throughout the study. The second measurement and CT scan were performed 10 min after the induction of pneumoperitoneum, with the patient still in the same position.

A frontal scout view covering the chest was performed before each spiral CT scan. Two spiral CT scans of the lungs (Philips Tomoscan SR 7000, Best, The Netherlands) were performed at end-expiration apnea with the patient in the supine position, before and after abdominal gas insufflation. This required an apnea period of approximately 12 s. A low-radiation-dose protocol according to the Early Lung Cancer Action Program was used. This includes besides pitch 2 (20-mm table movement for every 10-mm turn) also a low current–time product, 50 mA. The effective dose for the two scans was 1.2 mSv. Images of 10-mm thickness were reconstructed, covering the whole lung.

The measurement of the atelectasis area was made by manually drawing a encircled region of interest, containing pixels with values between −100 and +100 Hounsfield units (HU) (fig. 1). This definition of atelectasis was used according to suggestions from a previous study.12The pixel area was obtained in square millimeters for each 10-mm slice, which, multiplied by the slice thickness, gave the atelectasis volume. The change in atelectasis volume was calculated in percent change from the preinsufflation volume.

The overall lung volume was achieved by manually encircling the inner thoracic area excluding hilar structures in all slices from the apex to the costophrenic angles, containing pixels with values between −1,000 and 0 HU, and the volume from each slice was added.

To quantify the effect of pneumoperitoneum on regional volumes of gas and tissue, the lungs were divided into the upper and lower lobes. Because of the low-radiation-dose protocol used in this study, the lobes could be identified by knowledge of the bronchial divisions and intrapulmonary vessels. The right middle lobe was considered part of the right upper lobe.

To measure quantitatively the respective volumes of gas and tissue from the CT scans, a previously described method was used.13–15Each voxel of the CT scan is characterized by a CT number, which is numerically expressed in Hounsfield units. The scale ranges from +1,000 HU (bone) to 0 HU (water) and −1,000 HU (air). Density is the main determinant of the CT number. For example, a voxel with −500 HU consists of 50% gas and 50% tissue, and a voxel with −300 HU consists of 30% gas and 70% tissue. Tissue represents lung structures, extravascular water, and blood. All CT scans were reexamined with an interval of 100 HU, and the areas were measured. Then, it was manually calculated how much gas and tissue the areas contained. To find the total content, all volumes of gas and tissue, respectively, were then added. The overall volume of gas present in both lungs at end expiration was defined as functional residual capacity (FRC).

To describe the degree of aeration in the lung, the area of each slice from the apex to the costophrenic angles was drawn, and all slices were divided into four compartments according to their attenuation values. The different degrees of aeration were defined according to the following: overaeration = HU between −1,000 and −900,16normal aeration = HU between −900 and −500,17and poor aeration = HU between −500 and −100.17 

The distance between the top of the diaphragm and apex of the lung was measured on the CT images before and after abdominal gas insufflation. The sagittal and transverse diameters of the lung were measured from the CT scans at two levels, at Th6 and Th8, respectively. The sagittal length was measured 5 cm from the lateral border of the sternum at both the left and the right side of the lung (fig. 2).

After the first CT scan, a laparoscopic trocar was inserted under sterile conditions, and pneumoperitoneum was induced by insufflating carbon dioxide into the peritoneal cavity (Electronic Endoflator; Karl Storz, Tuttlingen, Germany). Intraabdominal pressure was monitored continuously through the laparoscope and maintained at 11–13 mmHg. No complications occurred during the investigation. When measurements were terminated, the patients were maintained under anesthesia and transported to a regular operating room.

Data are presented as mean ± SD. Statistical analysis was performed using the two-tailed paired Student t  test. A post hoc  test was performed by using the Tukey honest significant difference method. A P  value of less than 0.05 was considered significant.

After induction of pneumoperitoneum, there was a prompt increase in the mean peak airway pressure from 22 ± 2 to 25 ± 3 cm H²O (P < 0.01; table 1). Abdominal insufflation resulted in an increase of end-tidal carbon dioxide, from 27.0 ± 0.8 to 30.8 ± 2.2 mmHg although within the normal reference interval (P < 0.01; table 1). According to the protocol, respiratory rate, tidal volume, and fraction of inspired oxygen were not changed during the investigation. Heart rate did not change by the insufflation, although mean arterial pressure increased by 36% (P < 0.01; table 1).

In all seven patients, the volume of the atelectasis increased. The mean increase was from 49 ± 29 to 70 ± 33 cm³after abdominal gas insufflation (P < 0.01; table 2). This corresponded with a mean increase of 66% (range, 11–170%; fig. 3). In all patients, the atelectasis was dorsally localized, beginning at a level slightly above or at carina and increasing caudally (fig. 4).

Total end-expiratory lung volume decreased significantly in all seven patients after induction of pneumoperitoneum (P < 0.01; table 2). The mean decrease in percent was 11% (range, −23 to −4%).

Functional residual capacity decreased in all seven patients, from a mean value of 1.70 ± 0.55 l to 1.43 ± 0.48 l (P < 0.01; table 2). This corresponded to a mean decrease of 16% (range, −36 to −3%). The lung tissue volume also decreased in all but one patient from a mean value of 0.94 ± 0.12 l to 0.90 ± 0.10 l (P < 0.05; table 2). This corresponded with a mean decrease of 4% (range, −10 to +1%). When relating the changes in gas and tissue volumes to the change in total volume, the percentage of tissue of the total volume increased significantly after induction of pneumoperitoneum from 36 ± 6 to 39 ± 7% (P < 0.01), whereas the percentage of gas decreased significantly from 62 ± 7% to 59 ± 8% (P < 0.05; table 2).

After insufflation of carbon dioxide, an increase in mean percent tissue volume of the total lung volume was seen in both the upper (P < 0.01) and the lower lobes of the lungs (P < 0.05), whereas mean percent gas volume of the total lung volume was significantly decreased in both the upper and lower lobes (fig. 5and table 2).

Overaerated areas decreased significantly relative to the total lung volume (P < 0.05). Areas with poor aeration increased significantly (P < 0.05), whereas areas with normal aeration did not change significantly (fig. 6and table 3).

Induction of pneumoperitoneum resulted in a cranial displacement of the diaphragm between 1 and 3 cm (mean, 1.9 cm; P < 0.01). The reduction of the lung cephalocaudal dimension was equivalent on the right and the left sides in five of the seven patients. In two patients, the right side was reduced 1 cm more than the left side.

In all patients, the liver was compressed and dislocated caudally when gas was insufflated into the peritoneal cavity (fig. 4B).

The transverse diameter of the lung after insufflation did not significantly change compared with the diameter before insufflation, neither at Th6 nor at Th8. The sagittal diameters were significantly increased at both sides and at both Th6 and Th8 (table 4).

Several studies have shown that anesthesia per se  creates formation of atelectasis.1–4Although there have been many speculations of the impact of pneumoperitoneum on atelectasis formation, this relation has not been documented radiologically. The current study showed an increased volume of atelectasis in dependent lung regions and a cranial shift of the diaphragm after induction of carbon dioxide pneumoperitoneum. Total lung volume and FRC were reduced.

Previously, a relation between increased atelectatic area and increased shunt, resulting in a decreased arterial oxygenation, was observed5,7by CT in anesthetized patients. Surprisingly, previous results from our group have shown that pneumoperitoneum transiently reduces the shunt. We have shown this to be valid both indirectly by calculations according to the shunt formula8and directly by the multiple inert gas technique.9When the shunt was reduced, arterial oxygenation increased. An increase in arterial oxygenation during pneumoperitoneum has also been documented by others.10,11Therefore, we expected that pneumoperitoneum would reduce the formation of atelectatic area observed on CT. To the contrary, the atelectatic area increased in all patients.

It can only be speculated about why there is an improved arterial oxygenation when the atelectatic area is increased. There may be a better matching between ventilation and perfusion, in spite of the increase in atelectatic area. A significant portion of the shunt before pneumoperitoneum could occur in regions where, during pneumoperitoneum, the perfusion is decreased by the increased intraabdominal pressure transmitted to the thorax.

Supporting this theory is also the fact that, previously, when evaluating gas exchange by the multiple inert gas technique, we showed a decreased dispersion of blood flow (log of the standard deviation of the perfusion distribution) after induction of pneumoperitoneum, indicative of a better matching.9 

Unfortunately, in the current study, arterial oxygen tension was not measured. However, data from approximately 50 American Society of Anesthesiologists physical status I patients in previous studies8,9,18,19have shown increased oxygenation during pneumoperitoneum.

In the current study, propofol was used for anesthesia induction and maintenance. Besides ketamine, choice of anesthetic agent is most likely of no importance regarding the influence on the atelectasis development.6If hypoxic pulmonary vasoconstriction played a role for the current results, propofol is known not to interfere with this mechanism.20Because the current patients were normocapnic, increased carbon dioxide level did not influence the results.

There have been speculations that carbon dioxide insufflation changes the configuration of the thorax secondary to a restriction of the caudal movement of the thorax. Currently, the diaphragm was shifted 2 cm cranially after induction of pneumoperitoneum. This was an expected finding, but to our knowledge, this has not been radiologically visualized before. Theoretically, there could be a compensatory increase in sagittal and transverse diameters of the thorax. In the current study, the sagittal diameter at both Th6 and Th8 was increased, but there was no change in the transverse diameters of the thorax.

Currently, the cranial displacement of the diaphragm was accompanied by a reduction of the total lung volume and FRC. FRC is normally 2,300–2,600 ml in a 60-kg adult. FRC decreases by approximately 700–800 ml when changing from the upright or sitting position to the supine position. This decrease is further attenuated by anesthesia by 400–700 ml in adults, with only small alterations of thoracoabdominal dimensions.21In the current study, FRC decreased significantly by the induction of pneumoperitoneum; the mean decrease was 272 ml. The decrease of gas volume occurred in both the upper and lower lobes of the lung.

The clinical implications of the current findings regarding patients with lung disease remain to be elucidated. It has been shown previously that patients with chronic obstructive lung disease showed no or insignificant atelectasis during the first 45 min of anesthesia.22However, this may not be the case in the pneumoperitoneum situation. Then, increased atelectasis formation would add to negative influence of high airway pressure, decreased compliance, and increased arterial carbon dioxide level in this patient group.

In conclusion, the current study showed that pneumoperitoneum at an intraabdominal pressure level of 11– 13 mmHg displaces the diaphragm cranially and increases the volume of atelectasis in patients without heart or lung disease. It also showed an increase of tissue volume in both the upper and lower lobes of the lung. This indicates that the previously seen improvement in arterial oxygenation despite an increase in atelectasis volume could be due to an improved ventilation–perfusion matching during pneumoperitoneum.

The authors thank Anne-Marie Halén (Certified Registered Nurse of Anesthesia), Eva Olofsson (Certified Registered Nurse of Operating Room), and Paula Hellkvist (Certified Registered Nurse of Operating Room), and Anders Svensson (Radiology Nurse, Karolinska University Hospital, Huddinge, Sweden), for their skillful technical assistance. The authors also thank Alf Sollevi, M.D., Ph.D. (Professor, Department of Anesthesiology and Intensive Care, Karolinska University Hospital, Huddinge, Sweden), for valuable discussions.