Positive End-expiratory Pressure Improves Respiratory Function in Obese but not in Normal Subjects during Anesthesia and Paralysis 

The investigation was approved by the ethics committee of our institution, and informed consent was obtained from each subject preoperatively.

Nine normal subjects (five males) scheduled for abdominal surgery were included in the study. Their anthropometric and clinical characteristics are shown in table 1. Mean age and body mass index were 43.3 ± 15.2 yr (SD)(range, 24–66 yr) and 23.1 ± 1.2 kg/m²(range, 21.1–24.8 kg/m²), respectively. None had an history of smoking or clinical evidence of cardiopulmonary disease.

Nine morbidly obese patients (three males) admitted to the intensive care unit of our institution after abdominal surgery for ileojejunal bypass or gastric binding were included in the study (see table 1). Mean age and body mass index were 43.4 ± 14.4 yr (range, 19–68 yr) and 51.0 ± 8.2 kg/m²(range, 42.1–66.6 kg/m²), respectively. None of the patients had a history of smoking or clinical evidence of cardiopulmonary disease. Their preoperative pulmonary function tests, done in seated position, were as follows: Average forced vital capacity was 93.5 ± 8.8% of predicted, the forced expiratory volume in 1 s was 92.4 ± 8.6% of predicted, and the ratio of the latter to the former was 96.6 ± 8.8% of predicted. The average arterial blood-gas measurements during spontaneous ventilation in room air, before surgical intervention, included arterial oxygen tension (Pa^O2) of 88 ± 9 mmHg, arterial carbon dioxide tension (Pa^CO2) of 39±2 mmHg, and p  Ha of 7.41 ± 0.02.

In normal subjects and in obese patients anesthesia was induced with intravenous thiopental sodium (5–7 mg/kg). Muscle relaxation to facilitate endotracheal intubation was provided with succinylcholine (1 mg/kg), and maintenance of paralysis with pancuronium bromide. They were intubated with a cuffed endotracheal tube (Portex Ltd., Hythe, Kent, United Kingdom) and mechanically ventilated (tidal volume 8–12 ml/kg ideal body weight, and respiratory rate to maintain Pa^CO2between 35 and 45 mmHg) during the surgical procedure. Anesthesia was maintained with isofluorane 0.5–1.0% in 50/50% O²/N²O and boluses of fentanyl, if appropriate.

Normal subjects were studied in the recovery room about 20 min after the end of the surgical procedure; obese patients were immediately transported to the intensive care unit, still anesthetized and paralyzed. In obese patients, the entire protocol was performed in the intensive care unit after they reached respiratory and hemodynamic stability as judged by the physician in charge (137 ± 50 min after the end of the surgical procedure).

During the protocol, both normal subjects and obese patients were anesthetized with diazepam (0.1–0.2 mg/kg), paralyzed with pancuronium bromide (0.1–0.2 mg/kg), and mechanically ventilated with a ventilator (Siemens Servo 900C, Siemens-Elema, Solna, Sweden) in volume control mode (I:E ratio equal to 33%) with constant inspiratory flow (average tidal volume 0.683 ± 0.043 l [0.010 ± 0.001 l/kg ideal body weight] and 0.681 ± 0.040 l [0.010 ± 0.002 l/kg ideal body weight], inspiratory flow 0.488 ± 0.031 l/s and 0.480 ± 0.120 l/s, and respiratory rate 13.8 ± 0.7 breaths/min and 13.9±0.3 breaths/min in normal subjects and in obese patients, respectively). Inspiratory oxygen fraction was maintained equal to 50% during the entire protocol.

In all patients, we measured gas exchange; elastic and flow-resistive properties of the respiratory system, lung, and chest wall; and end-expiratory lung volume at two different PEEP levels (0 and 10 cm H²O) in normal subjects and in obese patients, applied in random order and maintained for at least 25–30 min. All measurements were made during the last 5 to 10 min of each level of PEEP. No previous volume history maneuver had been used in either group before starting the protocol. Other than changing PEEP level, baseline ventilatory settings were kept constant in each patient throughout the experiment. The entire protocol lasted about 1 h.

All the measurements were performed in supine position, without any abdominal bandage.

Intraabdominal pressure was measured using a transurethral bladder catheter. The technique has been previously reported and validated. 15Using a sterile technique, 100 ml normal saline was infused through the urinary catheter into the bladder. The catheter was then clamped and intraabdominal pressure recorded by a pressure transducer as mean pressure at end-expiration. The zero was set at the level of the pubis. The intraabdominal pressure was measured in only six out of nine normal subjects, those who had the bladder catheter already inserted for clinical purposes.

The end-expiratory lung volume was measured at end-expiration, at 0 cm H²O of PEEP, and at 10 cm H²O of PEEP using a closed-circuit helium dilution method. 16The expected end-expiratory lung volume was computed according to the formula derived from normal awake supine subjects. 17 

Airway pressure (P^ao) was measured proximal to the endotracheal tube by using polyethylene tubing (2 mm inner diameter, 120 cm long), connected to a Bentley Trantec pressure transducer (Irvine, CA). Gas flow was recorded by using a heated pneumotachograph (Hans Rudolph, Kansas City, MO) connected to an MP 45–1 differential pressure transducer (Validyne, Northridge, CA). Volume was obtained by digital integration of the flow signal.

Esophageal pressure (P^es) was measured from an esophageal balloon (Bicore CP-100, Irvine, CA) inflated with 0.5–1.0 ml of air, positioned at the lower third of the esophagus. The validity of P^eswas verified, according to Baydur et al.  18Because the patients were paralyzed, the airway and esophageal pressure were altered by compressing the thorax with the airways occluded at end-expiration. One could question the accuracy of P^esin obese patients while supine. Mediastinal organs may compress the esophagus and invalidate the translation into a pleural pressure. However, no alternative methods are available in humans, and this method has been previously adopted to partition respiratory mechanics in both awake 19and paralyzed obese patients in supine 5,19and prone positions. 20 

Both flow and pressure signals were recorded on a personal computer and processed via  an analog-to-digital converter (Colligo; Elekton, Milan, Italy) at a sample rate of 200 Hz and stored for subsequent computer analysis.

To determine the static elastance of total respiratory system, lung and chest wall, P^aoand P^eswere recorded during a 4- to 5-s airway occlusion at end-expiration and at end-inspiration. 21The pressure–volume (P-V) curves of the total respiratory system, lung, and chest wall were measured by inflating the respiratory system at different volumes (from 0.1 to 1.0 l, in 100-ml steps) and recording the corresponding P^aoand P^esduring airways occlusion at end-inspiration. 6,22End-expiratory volume corresponded to the elastic equilibrium volume in each patient, as evidenced by zero flow during an expiratory pause and absence of changes in P^aoafter airway occlusion.

Static elastance of the total respiratory system was computed as ΔP^ao/V^T, where ΔP^aois the difference between end-inspiratory and end-expiratory airway pressures and V^Tis the tidal volume. Static elastance of the chest wall was computed as ΔP^es/V^T, where ΔP^esis the difference between end-inspiratory and end-expiratory esophageal pressures. Static lung elastance was calculated as respiratory system elastance minus chest wall elastance.

From the P-V curve we derived the following parameters:23 

1. Starting elastance:  This was computed as the ratio between the pressure at 100 ml inflation and the corresponding volume. The starting elastance at PEEP 0 cm H²O likely reflects the mechanical characteristics of the respiratory system at end-expiration. In fact, during the measurement of the starting elastance it is unlikely that any significant alveolar recruitment or small airway reopening occur, because of the low pressures used

2. Inflation elastance:  This was computed as the slope of the P-V curve during inflation in its most linear segment; to do so we excluded by visual inspection the points included in the inflection region. Throughout the remaining points of the inflation limb we performed a linear regression. The reciprocal of the slope of the linear regression was defined as the inflation elastance.

3. Lower inflection point:  This was computed as the pressure corresponding to the intersection between the starting and inflation elastance obtained by a computerized system.

4. The ratio of inflation elastance to starting elastance was computed as a quantitative estimate of the curvilinearity of the P-V curve.

At PEEP these computations were performed as a change from the end-expiratory pressure.

We first computed the difference of gas volume measured with helium at 10 cm H²O and 0 cm H²O pressure (actual Δ gas volume). We then computed on the P-V curve, starting at PEEP 0 cm H²O, the gas volume expected at 10 cm H²O. We defined as recruitment the difference between the actual Δ gas volume and the expected Δ gas volume.

Maximum and minimum resistances of the respiratory system were computed from P^aoas (P^max′− P²)/Vⁱ′ and (P^max′− P¹)/Vⁱ′, where P^max′ is the maximal pressure value after occlusion corrected for the tube resistance (which was determined experimentally), P¹is the pressure recorded after the immediate drop from Pmax, P²is the plateau pressure, and Vⁱ′ is the flow immediately preceding the occlusion. The minimum resistance represents the “ohmic” resistive component of the respiratory system, and the maximum resistance includes the minimum plus the “additional” respiratory resistance caused by stress relaxation or time constant inequalities within the respiratory system tissues. There was no appreciable decrease in P^esimmediately after the occlusion (i.e. , P¹in the esophageal tracings was not identifiable); the minimum resistance of the respiratory system reflects essentially airway resistance; and minimum chest wall resistance can be considered negligible. 21As a consequence, maximum chest wall resistance results entirely from to the viscoelastic properties of the chest wall tissues. Additional resistance of the lung was obtained as additional respiratory resistance minus additional chest wall resistance; the sum of minimum lung resistance and additional lung resistance gives the maximum lung resistance. Additional lung and chest wall resistances result from stress relaxation or time constant inequalities within the lung and chest wall, respectively.

Arterial blood samples from the radial artery were analyzed for p  Ha, Pa^O2, and Pa^CO2by a blood-gas analyzer (IL 1312 BGM; Instrumentation Laboratory Company, Lexington, MA). Expired gases were collected over a 3-min period in a Douglas bag at the same time as blood-gas samples. Expired gas was analyzed for mixed expired carbon dioxide concentration (Normocap, Datex, Finland). The physiologic dead-space fraction (V^D/V^T) was computed according to the following formula: where Pe^CO2is the mixed expired carbon dioxide partial pressure.

Values are expressed as mean ± SD. The mean value of three breaths was used for each variable and for each experimental condition. Comparisons between normal subjects and obese patients were performed using the unpaired Student t  test at both PEEP 0 and 10 cm H²O. Values obtained at PEEP 10 cm H²O were compared with PEEP 0 cm H²O using the two-way analysis of variance. Regression analysis was performed with the least-squares method. 24 

Table 2summarizes some of the most relevant mechanical variables of the normal subjects and obese patients after surgery and their changes with PEEP. It is noteworthy that although in normal subjects the measured end-expiratory lung volume at PEEP 0 cm H²O was similar to the end-expiratory lung volume expected (2.154 ± 0.584 vs.  2.271 ± 0.454 l, P  not significant), in obese patients it was only 28% of expected (0.587 ± 0.175 vs.  2.209 ± 0.596 l, respectively;P < 0.01).

At PEEP 0 cm H²O the elastances of the respiratory system, lung, and chest wall were significantly higher in obese patients compared with normal subjects. In obese patients the elastance of the total respiratory system significantly decreased with PEEP, because of a similar reduction in lung and chest wall elastance. In normal subjects, however, PEEP produced a slight reduction in lung elastance, without affecting total respiratory system and chest wall elastance.

Finally, it is worth noting that, on average, the intraabdominal pressure was almost twofold greater in obese patients compared with normal subjects (18.8 ± 7.8 vs.  9.0 ± 2.4 cm H²O, P < 0.01). However, unexpectedly, the two most severe obese patients had intraabdominal pressures in the range observed in normal subjects, (patients 13 and 17, intraabdominal pressure = 9 and 8 cm H²O, respectively).

In figure 1, the P-V curves of the total respiratory system, lung, and chest wall, are reported for both normal and obese patients. In normal subjects the P-V curves at PEEP 0 and 10 cm H²O roughly follow the same pattern. In the obese patients, however, the P-V curves at 10 cm H²O of PEEP are shifted upward and to the left, suggesting alveolar recruitment.

The quantitative analysis of P-V curves is reported in table 3. As shown, the inflation and starting elastances derived from the P-V curve behave as the “cord elastances”(i.e. , the elastances measured as Δpressure/Δvolume between end-inspiration and end-expiration and reported in table 2). Two additional pieces of information, however, may be derived from the P-V curve. First, the ratio of inflation elastance to starting elastances of the total respiratory system, lung, and chest wall, which estimates the curvilinearity of the P-V curve, is similar in obese patients and normal subjects at PEEP 0 cm H²O, and in both populations, with increasing PEEP, the ratio tends to increase, indicating a more linear P-V curve at 10 cm H²O of PEEP. Finally, the lower inflection point (P^flex) of the total respiratory system was significantly higher at PEEP 0 cm H²O in the obese patients and decreased with PEEP. However, at 10 cm H²O PEEP, the obese patients and normal subjects showed similar P^flexvalues.

The resistances of the total respiratory system, lung, and chest wall in both normal and obese patients at PEEP 0 and 10 cm H²O are shown in table 4. At PEEP 0 cm H²O, maximum and additional resistance of the total respiratory system were significantly higher in obese patients compared with normal subjects. This increase resulted mainly from maximum and additional lung resistance. On the other hand, airway and chest wall resistances were similar in normal subjects and in obese patients. PEEP significantly decreased maximum resistance of the total respiratory system mainly by reducing the lung component, but the effect was more evident in obese patients than in normal subjects. The decrease in maximum lung resistance with PEEP was prevalent in the airway resistance component both in normal subjects and in obese patients.

As shown in table 5, despite similar minute ventilation and identical inspired oxygen fraction, the obese patients had significantly lower Pa^O2values, higher alveolar–arterial oxygen difference, higher Pa^CO2values, and higher physiologic dead-space fraction compared with normal subjects. The normal subjects did not show any significant change in gas exchange on increase of PEEP; the obese patients had significantly increased Pa^O2values and decreased alveolar–arterial oxygen differences. However, the differences in Pa^O2, alveolar–arterial oxygen difference, Pa^CO2, and dead space remained significant between obese and normal subjects also at 10 cm H²O of PEEP. It is noteworthy that, although the average changes of Pa^O2in normal subjects were not significant with increasing PEEP, in seven of nine normal patients the Pa^O2actually decreased; Pa^O2decreased in only one of nine obese patients.

The recruitment was significantly higher in obese patients compared with normal subjects (0.153 ± 0.168 and −0.060 ± 0.246 l, respectively;P < 0.05). In the whole study population (normal and obese) the increase of Pa^O2with PEEP was significantly correlated with recruitment (r = 0.81, P < 0.01;fig. 2) as the alveolar–arterial oxygen difference (r = 0.74, P < 0.01). As shown, the normal subjects are mostly located in the nonrecruitment area (and did not have increases in Pa^O2), except for one normal subject (patient 8) who had a consistent recruitment and Pa^O2increase and experienced the same effects as the obese patients. It is noteworthy that the statistical significance of figure 2results entirely from the obese patients (for this population alone, r = 0.95, P < 0.01); the marginal recruitment, or even derecruitment, observed in the normal population considered alone was not correlated significantly with the Pa^O2changes.

In the whole population, at PEEP 0 cm H²O, we found a significant correlation between the body mass index and decreased end-expiratory lung volume (r = 0.83, P < 0.01), increased elastance of the respiratory system (r = 0.86, P < 0.01), decreased Pa^O2(r = 0.72, P < 0.01), and increased Pa^CO2(r = 0.62, P < 0.01). The chest wall elastance was significantly correlated with decreased end-expiratory lung volume (r = 0.68, P < 0.01), decreased Pa^O2(r = 0.62, P < 0.01), and increased Pa^CO2(r = 0.77, P < 0.01).

As the chest wall elastance appears to be a function of intraabdominal pressure (r = 0.94, P < 0.01;fig. 3), it is tempting to hypothesize the following sequence of events: Increasing body mass leads to increased intraabdominal pressure, which in turn, by decreasing the chest wall elastance, causes lung volume reduction and gas-exchange deterioration. However, we did not find any significant relationship between body mass index and intraabdominal pressure, and this suggests a more complex relationship among body mass, intraabdominal pressure, respiratory mechanics, and gas-exchange deterioration.

In this study, we found remarkable differences at PEEP 0 cm H²O in lung volumes, respiratory mechanics, and gas exchange in the early postoperative period, during anesthesia and paralysis, between normal and morbidly obese patients.

Moreover, we found that increasing PEEP to 10 cm H²O led to a significant improvement of respiratory mechanics and gas exchange in the obese patients but not in normal subjects. We discuss the possible mechanisms underlying the differences we found between normal subjects and obese patients at PEEP 0 cm H²O, and their different responses to PEEP 10 cm H²O.

Reportedly, the end-expiratory lung volumes of obese patients before anesthesia and paralysis, as well as lung and chest wall elastances, are similar to normal. 25Damia et al. , 16using the same method of measurement we used, found normal values of end-expiratory lung volume in a population of morbidly obese patients before anesthesia. Indeed, despite the fact that preanesthetic end-expiratory lung volume values of our obese patients were likely similar to normal, as also suggested by the preoperative pulmonary function tests, the end-expiratory lung volume decrease during anesthesia and paralysis was dramatically different in obese patients compared with normal subjects. It is tempting to speculate that the cause of the end-expiratory lung volume decrease is the increased abdominal pressure unopposed by the diaphragm tone during anesthesia and paralysis. 26,27In fact, we found that, on average, the obese patients were characterized by an increased intraabdominal pressure. Moreover, increased abdominal pressure was associated with a reduction in end-expiratory lung volume and increased chest wall and respiratory system elastances. It is possible that the increased chest wall elastance led to a decrease of transpulmonary pressure, with consequent lung-size reduction or collapse. The importance of intraabdominal pressure is underlined by the findings of Damia et al. , 16who observed that, after opening of the abdominal wall, the lung volume in obese patients almost returned to the preanesthesia values.

However, the increased intraabdominal pressure may not entirely explain the end-expiratory lung volume reduction. In fact, we were not able to find any significant relationship between intraabdominal pressure and the body mass index; in the two patients with the highest body mass indices the intraabdominal pressure and the chest wall elastance were near normal, and the lung volume was greatly decreased. Moreover, in a previous study, 28we found that increasing the intraabdominal pressure up to 20 cm H²O during laparoscopy in normal subjects during anesthesia and paralysis resulted only in a moderate decrease in end-expiratory lung volume (about 20%). These findings suggest that intraabdominal pressure is not the unique cause of end-expiratory lung volume reduction, and other factors are likely involved, such as the blood shift from abdomen to thorax 29and the distortion of the rib cage, without diaphragm displacement. 30,31Indeed, the exact mechanism leading to end-expiratory lung volume decrease remains an open question.

Whatever the cause of end-expiratory lung volume reduction, the respiratory mechanics are more affected in obese patients than in normal subjects. In obese patients lung and chest wall elastances are twofold greater than in normal patients, and the P-V curves show a greater inflection point, resulting from both the lung and chest wall components. It is known that, during anesthesia and paralysis, even normal subjects increase elastance, and reportedly 32the major component of the changes in elastance is the lung rather than the chest wall. Several mechanisms have been proposed to explain the elastance changes: Besides the reduction of end-expiratory lung volume, other factors may be involved, such as the distortion of rib cage (and lung), the cephalad shift of the diaphragm, surfactant alteration, blood shift from abdomen to thorax, or a combination of these. 33–35 

Reduced end-expiratory lung volume and increased elastance are associated with a more severe deterioration in gas exchange in obese patients compared with normal subjects, who show a milder abnormality in alveolar–arterial oxygen difference. Several mechanisms have been proposed to explain the gas-exchange deterioration during anesthesia in normal subjects, such as small-airways collapse 36and altered ventilation distribution with consequent ventilation–perfusion (V̇^A/Q̇) mismatch, 7and compression atelectasis with alveolar collapse and true shunt. 37It is likely that the mechanisms leading to the deterioration of gas exchange in normal subjects are operating to a higher degree in the obese patients. In fact, in the obese patients, the lower lung volume should enhance the small-airways collapse, with higher degree of V̇^A/Q̇ mismatch, as also suggested by the increased Pa^CO2and dead space compared with normal. Moreover, it has been shown that the compression atelectasis increases with body size, 38and although we lack direct evidence, it is likely that compression atelectasis and true shunt are higher in obese patients compared with normal subjects.

Raising the PEEP in normal subjects to 10 cm H²O neither increased Pa^O2nor decreased the alveolar–arterial oxygen difference; in the obese patients, it did so. As we did not measure the hemodynamics or the V̇^A/Q̇ distribution, any explanation of our findings is only speculative. The role of PEEP in anesthesia remains controversial, as some authors have found a variable response to oxygenation in different subsets of patients and clinical conditions. 7,12,13,39It is well known that increasing PEEP may both improve or deteriorate Pa^O2. The resolution of atelectasis, if present, should increase Pa^O2, and the prevention of small-airways closure should improve the V^A/Q matching. Furthermore PEEP, increasing lung volume, should induce a more even distribution of ventilation, with possible improvement of V̇^A/Q̇. On the other hand, increasing PEEP may induce overstretching of pulmonary units and cardiac output decrease or redistribution, with possible negative effects on V̇^A/Q̇ and shunt, and these hemodynamic consequences may offset the respiratory advantages. 40,41It is then likely that the response to PEEP depends on the balance between positive and negative effects in any given patient.

The failure to improve oxygenation in normal subjects has been attributed to the contemporary decrease of cardiac output with PEEP. 7In our study, it appears that the positive effect of PEEP on the obese patients outweighs its negative effects, 40,41but in normal subjects the positive and negative effect of PEEP seem balanced, resulting in unchanged oxygenation.

Although we do not have the “gold standard” for the recruitment evaluation (i.e. , computed tomographic scan), several indirect findings suggest that the collapse is greater in obese patients than in normal subjects. In fact, the obese patients have lower lung volume, greater elastance, and evident lower inflection point. Consequently, the PEEP in obese patients induces recruitment, as suggested by the upward shift of the P-V curve and the decrease of inflection point, and the higher the recruitment, the higher the increase in oxygenation, as shown in figure 2. We thus believe that obesity leads to conditions under which the positive effects of PEEP outweigh its potential negative effects.

From the body of these data it is tempting to present an overall interpretation of the effects on the obese patients during anesthesia and paralysis. It appears that anesthesia and paralysis cause the reduction of the lung volume through a continuum related to the body mass index. 42The reduction of lung volume is associated to an increase of the elastance of the respiratory system, equally divided between its chest wall and lung components. From our data it seems that increased intraabdominal pressure may play an important (although not unique) role in increasing the chest wall elastance and decreasing the end-expiratory lung volume. The decrease of oxygenation is associated with the increase of body mass index and lung volume reduction, suggesting that a relevant lung collapse is likely present in the obese patients.

Applying 10 cm H²O of PEEP increases oxygenation in obese patients but not in normal subjects, and the oxygenation improvement is related to the amount of alveolar recruitment. The “opening” of collapsed units is a function of transmural pressure, which is mainly related to plateau pressure, that is, tidal volume. In this study we used relatively low tidal volume, and the plateau pressure we reached starting from 10 cm H²O of PEEP averaged 22.1 ± 1.9 and 20.7 ± 2.5 cm H²O in obese patients and in normal subjects, respectively. It is possible that we did not fully open the lung at end-inspiration, because the transmural pressure we reached could be lower than necessary. 3 

We did not investigate the effects of different tidal volumes on respiratory mechanics and oxygenation, but considering the underlying pathophysiology, it is possible that higher tidal volumes may be beneficial in obese patients. 40,43,44Indeed, further studies are warranted to define the optimal levels of tidal volume and PEEP to open up and keep open the lung of obese patients, as well as to define at which level of body mass index PEEP or larger tidal volume is effective to maintain normal respiratory mechanics and oxygenation.