Obesity Surgery

, Volume 18, Issue 3, pp 256–263

Alveolar-Membrane Diffusing Capacity Improves in the Morbidly Obese after Bariatric Surgery

Authors

    • Department of Obstetrics, Gynecology and Women’s Health, School of Medicine, Saint Mary’s Health CenterSaint Louis University
    • Department of Pharmacological and Physiological Science, School of MedicineSaint Louis University
  • Do Jun Kim
    • Department of Obstetrics, Gynecology and Women’s Health, School of Medicine, Saint Mary’s Health CenterSaint Louis University
  • Jean-Loup Sylvestre
    • Department of Surgery, Bariatric Clinic, Royal Victoria HospitalMcGill University Health Center
  • Nicolas V. Christou
    • Department of Surgery, Bariatric Clinic, Royal Victoria HospitalMcGill University Health Center
Research Article

DOI: 10.1007/s11695-007-9294-9

Cite this article as:
Zavorsky, G.S., Kim, D.J., Sylvestre, J. et al. OBES SURG (2008) 18: 256. doi:10.1007/s11695-007-9294-9
  • 115 Views

Abstract

Background

Morbidly obese individuals may have impaired alveolar-membrane diffusing capacity (DmCO). The purpose of this study was to measure pulmonary diffusing capacity for NO (DLNO) as an index of DmCO pre- and postbariatric surgery in the morbidly obese.

Methods

Twenty-one patients [age = 40 ± 9 years, body mass index (BMI) = 48.5 ± 7.2 kg/m2] with an excess weight of 72 ± 17 kg scheduled for bariatric surgery were recruited. Pulmonary function and arterial blood-gases were measured pre- and postsurgery.

Results

DmCO was 88 ± 23% of predicted before surgery (p < 0.05). There was loss in BMI and excess weight of 7.7 ± 2.0 kg/m2 and 31 ± 8%, respectively. Because DmCO = DLNO/2.42, the increase in DLNO postsurgery resulted in a normalization of the predicted DmCO to 97 ± 29% predicted, or an improvement of DLNO by 11 ± 18 (95% CI = 3.5, 19.1; p = 0.01) milliliters per minute per millimeter of mercury without any improvement in DLCO. The DLNO/DLCO ratio and alveolar volume both increased, respectively (p < 0.05), and pulmonary capillary blood volume to DmCO ratio decreased postsurgery (p < 0.01). Multiple linear regression revealed that the change in DLNO was most strongly associated with changes in alveolar volume and the waist-to-hip ratio (adjusted r2 = 0.76; p < 0.001) and was not related to the reduction in the alveolar-to-arterial PO2 difference.

Conclusion

Alveolar-membrane diffusion normalizes within 10 weeks after bariatric surgery. This is likely due to the increase in alveolar volume from the reduction in the waist-to-hip ratio.

Keywords

Nitric oxide diffusing capacityObesityWeight lossPulmonary functionSurgeryBariatric

Introduction

Pulmonary gas exchange has been shown to be affected by morbid obesity [body mass index (BMI) > 40 kg/m2] [1]. Specifically, the impairment in pulmonary gas exchange, the alveolar–arterial PO2 difference (AaDO2), has been associated with the waist-to-hip ratio in the morbidly obese, meaning that those with a higher AaDO2 have a larger waist circumference and a larger waist-to-hip ratio [2]. If the pulmonary gas exchange impairment can be related to the waist-to-hip ratio in the morbidly obese, then pulmonary diffusing capacity, alveolar-membrane diffusing capacity, and pulmonary capillary blood volume could also be impaired compared to predicted norms.

Over the past 15 years, the use of nitric oxide (NO) in conjunction with carbon monoxide (CO) has been used to obtain alveolar-membrane diffusing capacity for carbon monoxide (DmCO), and through interpolation, pulmonary capillary blood volume (Vc), such that pulmonary diffusing capacity for NO (DLNO) and CO (DLCO), DmCO, and Vc are obtained in a one-step process. This process has several advantages compared to the two-step Roughton and Forster method [3], which is described elsewhere [4, 5].

Examining changes in alveolar-membrane conductance with weight loss using DLNO has not yet been studied to date. Therefore, the purpose of this study was to examine the change in DLNO in morbidly obese patients’ pre- and 10 weeks postbariatric surgery and to determine which factor(s) is (are) associated with the change in DLNO. As the pulmonary gas exchange impairment has been shown to be related to the waist-to-hip ratio [2], our present hypothesis was that DLNO would improve as a result of bariatric surgery, and the improvement would be due to the increase in alveolar volume caused by the decrease in the waist-to-hip ratio.

Methods

Subjects

Morbidly obese subjects (BMI > 40 kg/m2) were recruited from the bariatric clinic. Each subject was required to participate in two testing sessions, one pre- and one postsurgery testing session, about 2 months apart. The tests included body composition, cumulative blood counts (CBCs), pulmonary function (spirometry, lung diffusion measurements) arterial blood sampling, and measurement of resting metabolic rate. Spirometric function included measurements of slow and forced vital capacity (SVC and FVC), forced expiratory volume in 1 s (FEV1), peak expiratory flow rate (PEF), and forced expiratory flow rate over the middle half of expiration (FEF25–75). Excluded from the population of the morbidly obese were individuals with (1) BMI ≥ 70 kg/m2; (2) respiratory, renal, or hepatic failure; (3) metastatic disease; (4) senility, Alzheimer’s disease, or other dementias; and (5) inability to compre-hend the instructions during tests. All subjects signed an informed consent form. This study was approved by the Institutional Review Board of the University Health Centre.

Body Composition, Venous Blood Samples, and Arterial Cannulation

Height, weight, spirometric function, and body composition were assessed pre- and postsurgery. Lean and fat masses were measured from an 8 polar bioelectrical impedance device that has been validated for the morbidly obese [6]. The excess weight was estimated according to a formula [7] and is based on the Metropolitan Tables for middle-frame individuals. The percent excess weight loss was calculated as 100 × [(W0 − W1)/EW0], where W0 is the weight in kilograms at the time of surgery, W1 is the weight in kilograms at the last follow-up, and EW0 is the excess weight at the time of surgery.

Venous blood hematocrit, hemoglobin, white blood cell, and red blood cell were measured in each subject. Then, heart rate, oxygen consumption, and arterial blood-gases were measured over 5 min of rest, sitting upright on a chair, and corrected for changes in arterial blood temperature. The methodology of the set-up for arterial sampling and withdrawal are located elsewhere [2]. Arterial blood PO2, and PCO2, and pH were measured directly via an ABL725 Blood-Gas Analyzer (Radiometer, Copenhagen, Denmark). Arterial oxygen saturation (%SaO2) was also directly measured on the ABL725 analyzer using multiwavelength oximetry. The ideal alveolar gas equation was used to calculate alveolar PO2 (PAO2) such that the AaDO2 could be calculated [8]. Oxygen consumption at rest was assessed with a metabolic cart (model VMax 229LV; Sensormedics, Yorba Linda, CA, USA) and heart rate was assessed with a three lead ECG (Cardiocap/5, Datex Ohmeda, Louisville, CO, USA).

Single-Breath DNLO–DLCO Apparatus and Technique

Volume and gas calibration of the lung diffusion system (Hyp’Air, Medisoft®, Dinant, Belgium) were performed just prior to each testing session. The concentrations of inspiratory gases, types of gas analyzers used for measuring inspired and expired gas mixtures, methodology of the technique, and apparatus including deadspace washout volume and sampling volume are described elsewhere [4, 9]. Inspired volume (IV) was measured, corrected for instrument and anatomical deadspace, and converted to standard temperature and pressure, dry conditions. Breath-holding time was calculated using the method of Jones and Meade [10].

Calculation of Diffusion Capacities and Components

Diffusion capacities for NO and CO were calculated simultaneously from the exponential disappearance rate of each gas with respect to helium using the Jones–Meade method [10]. The formulae for calculating DLCO are from the 2005 American Thoracic Society (ATS) and European Respiratory Society (ERS) Guidelines [11]. The patient’s actual hemoglobin (Hb) concentration and PAO2 of 100 mm Hg was used to calculate the ΘCO for each subject by inserting the values into following formula by Roughton and Forster 3: \( \frac{1} {{Θ {\text{CO}}}} = {\left( {0.73 + 0.0058 \times {\text{PO}}_{2} } \right)} \times \frac{{14.6}} {{{\left[ {{\text{Hb}}} \right]}}} \).

Therefore, the 1/ΘCO ranged from 1.188 to 1.678 and ΘCO ranged from 0.596 to 0.842. The DLCO was corrected to the patient’s actual hemoglobin concentration using the formulas from the 2005 ATS and ERS Guidelines [11]. We have assumed, as most workers do, that ÈNO in vivo [5, 9, 1215], although not in vitro [16], is infinity. The likely reason for this discrepancy is facilitated diffusion and/or lack of a stagnant layer in vivo [17]. Therefore, the DLNO-to-DmCO ratios of 2.42 [5, 14, 15] and 1.97 [9, 12, 13, 18, 19] were calculated as the theoretical ratio of DLNO to DmCO during single-breath maneuvers because those ratios have both been used in the past. The ratio of 1.97 is based on the solubility of CO and NO in plasma, whereas a higher ratio of 2.42 has been determined recently during rebreathing maneuvers at rest and during exercise [5, 15]. The higher ratio of 2.42 may be due to the larger-than-assumed NO solubility in plasma, as well as NO-facilitated diffusion [5]. Because it has been recently established that using the ratio of 2.42 gives a combined better estimate of Dm and Vc than the ratio of 1.97 [4, 20] when compared to established norms elsewhere [21], the ratio of 2.42 was used for the calculation of alveolar-membrane diffusing capacity and pulmonary capillary blood volume in this study.

A breath-holding time of 5 s for the maneuver was chosen because it has been shown that DLCO values are not different with a 3- or 5-s breath hold compared to a 7- and 10-s breath hold [22]. As the recent ATS and ERS guidelines recommend a minimum of 4 min between DLCO measurements [11], subjects performed the single-breath maneuver with a minimum of 4 min and 30 s between tests. The average of two properly performed maneuvers where DLNO and DLCO were within 17 and 3 mL·min−1·mm Hg−1 of each other, respectively, and with an inspired volume of at least 85% of the individual’s measured FVC was used for further calculation for each variable at each testing session [4].

Statistical Analyses

A paired t test was used to compare predicted spirometry and diffusing capacity parameters with the measured values. A paired t test was used to compare the presurgical values of anthropometric, CBCs, arterial blood-gases, and diffusing capacity parameters to the postsurgical values. The 95% confidence interval for the mean change between presurgical and postsurgical values was calculated for all parameters (the mean difference ± 1.96 × standard error of the mean).

Stepwise multiple linear regressions were performed to discover the factors that are associated with the change in DLNO from bariatric surgery. The independent variables (predictors) were gender and changes in AaDO2, %SaO2, alveolar volume, FVC, waist-to-hip ratio, and weight. The dependent variable was the change in DLNO. Because BMI is a function of weight and height together, and because height is associated with FVC, BMI was not used in the regression model. Once the predictors were placed in the stepwise algorithm, a model was built to assess the contribution of each one of them from the significance value of each t test for the predictor (entry criterion: p = 0.05). This significance value was compared against the removal criterion (p = 0.10) and removed if it did not make a significant contribution to the model. After a variable is entered, then all previous variables that were already in the model were examined to see if any of them meet the criteria for removal. It is possible for a variable that has initially met the entry criterion for the model to be removed at a later step. (This happens when independent variables are correlated with one another.) The statistical program used for all analyses was SPSS version 14.0 (SPSS, Chicago, IL, USA).

Results

Twenty-five community-dwelling, ambulatory, morbidly obese individuals scheduled for laparoscopic gastric bypass were recruited to participate. The surgery was performed under general anesthesia and none were ventilated postsurgery. One subject died 2 days after surgery and three other subjects refused to come in for their postsurgery testing, so the data are presented for 21 subjects. Nine of those subjects had metabolic syndrome prior to surgery, as defined by having at least three of five clinical markers [23]. The time between the presurgical and postsurgical measurement was 71 ± 22 days or 10.1 ± 3.1 weeks. In that time, the subjects lost 31% of their excess weight, which was equivalent to 22 kg of body mass or a reduction in BMI by 7.7 kg/m2 (Table 1).
Table 1

Changes in anthropometric parameters with weight loss

n = 21 (women = 14; men = 7)

Presurgery

Postsurgery

Change

95% confidence interval for the change

Mean (SD)

Mean (SD)

Mean (SD)

Anthropometric parameters

Age (years)

40 (9)

Height (cm)

167 (0.09)

Weight (kg)

135.3 (18.9)

113.6 (16.0)

−21.7 (6.1)a

−24.3, −19.1

Excess weight (kg)

71.9 (17.2)

50.3 (15.5)

−21.7 (6.1)a

−24.3, −19.1

Percent excess weight loss

31 (8)a

BMI (kg/m2)

48.5 (7.2)

40.8 (6.5)

−7.7 (2.0)a

−8.6, −6.8

Body surface area (m2)

2.4 (0.2)

2.2 (0.2)

−0.2 (0.1)a

−0.24, −0.16

Waist circumference (cm)

133 (15)

119 (14)

−15 (8)a

−18.4, −11.6

Hip circumference (cm)

139 (14)

129 (14)

−10 (5)a

−12.1, −7.9

Waist-to-hip ratio

0.97 (0.11)

0.92 (0.10)

−0.04 (0.07)a

−0.07, −0.01

Lean tissue mass (kg)

67.5 (12.4)

60.5 (11.7)

−7.0 (2.0)a

−7.9, −6.1

Fat mass (kg)

67.8 (12.7)

53.1 (13.1)

−14.7 (5.3)a

−17.0, −12.4

Body fat (%)

50 (6)

46 (8)

−4 (3)a

−4.8, −2.4

aThe change between pre- and postsurgery is significant (p < 0.05). Body surface area (BSA) was calculated as 0.0097 (height in cm+weight in kg) − 0.545.

There was a significant change in venous blood parameters pre- and postsurgery (Table 2). A reduction in hemoglobin, hematocrit, and total white and red blood cell count were shown. Pulmonary gas exchange improved with weight loss. This was evidenced by the decrease in AaDO2 and the increase in %SaO2 with weight loss (Table 3). In fact, 61% of the increase in %SaO2 was accounted for by the decrease in AaDO2 (r = −0.78, p < 0.05). Heart rate and oxygen consumption (L/min) at rest also decreased with weight loss (Table 3).
Table 2

Changes in venous blood parameters with weight loss

n = 21 (women = 14; men = 7)

Presurgery

Postsurgery

Change

95% confidence interval for the change

Mean (SD)

Mean (SD)

Mean (SD)

Venous blood parameters

Hemoglobin (g/L)

139 (11.7)

136 (11.1)

−3 (6.4)a

−5.7, −0.3

Hematocrit

0.417 (0.036)

0.406 (0.033)

−0.011 (0.018)a

−0.019, −0.003

Red blood cell count (T/L)

4.84 (0.43)

4.71 (0.41)

−0.13 (0.21)a

−0.22, −0.04

White blood cell count (G/L)

8.48 (1.89)

6.72 (1.23)

−1.76 (1.41)a

−2.36, −1.16

aThe change between pre- and postsurgery is significant (p < 0.05)

Table 3

Changes arterial blood-gas and metabolic parameters at rest with weight loss

n = 21 (women = 14; men = 7)

Presurgery

Postsurgery

Change

95% confidence interval for the change

Mean (SD)

Mean (SD)

Mean (SD)

Arterial blood-gases

PaO2 (mm Hg)

86 (10)

89 (5)

2 (8)

−1.4, 5.4

PAO2 (mm Hg)

106 (4)

103 (5)

−3 (4)a

−4.7, −1.3

AaDO2 (mm Hg)

20 (9)

14 (5)

−5 (7)a

−8.0, −2.0

PaCO2 (mm Hg)

37 (3)

36 (2)

−1 (3)a

−2.3, 0.3

%SaO2

97.1 (1.1)

97.6 (0.4)

0.4 (0.9)a

0.0, 0.8

pH

7.42 (0.04)

7.44 (0.02)

0.02 (0.04)a

0.0, 0.04

Arterial blood temperature (°C)

36.6 (0.3)

36.3 (0.3)

−0.2 (0.4)a

−0.4, −0.03

Metabolic variables

VO2 (L/min)

0.34 (0.08)

0.28 (0.07)

−0.06 (0.08)a

−0.1, −0.03

VO2 (mL kg−1 min−1)

2.6 (0.6)

2.5 (0.6)

0.0 (0.6)

−0.3, 0.3

VE (L/min)

11.1 (1.9)

9.4 (1.7)

−1.7 (2.4)a

−2.7, −0.7

RER

0.83 (0.06)

0.74 (0.06)

−0.09 (0.07)a

−0.12, −0.06

HR (beats/min)

85 (17)

72 (15)

−14 (17)a

−21, −7

PaO2 = arterial PO2, PAO2 = alveolar PO2, AaDO2 = alveolar to arterial PO2 difference, %SaO2 = arterial oxyhemoglobin saturation measured via co-oximetry, VO2 = oxygen consumption, VE = minute ventilation, RER = respiratory exchange ratio, HR = heart rate.

aThe change between pre- and postsurgery is significant (p < 0.05)

Four morbidly obese subjects had abnormal spirometry function presurgery, as their FEV1 or FVC was below the lower limits of normal (LLN) 24. As a group, the spirometric function was different compared to predicted values (Table 4). There was an improvement in spirometric function from the surgery as the FEV1 and FVC increased by ∼7 to 9%. The four morbidly obese subjects who had their FEV1 or FVC below the LLN presurgery now all had values above the LLN.
Table 4

Changes in spirometry with weight loss

n = 21 (women = 14; men = 7)

Presurgery

Percent predicted presurgery

Postsurgery

Change

95% confidence interval for the mean change

Mean (SD)

Mean (SD)

Mean (SD)

Mean (SD)

Spirometry

FEV1 (L)

3.06 (0.68)

91 (11)a

3.34 (0.71)

0.28 (0.24)b

0.18, 0.38

FVC (L)

3.93 (0.82)

94 (10)a

4.21 (0.87)

0.29 (0.17)b

0.22, 0.36

SVC (L)

4.04 (0.76)

4.26 (0.87)

0.22 (0.29)b

0.10, 0.34

FEV1/FVC (%)

78 (6)

79 (3)

1 (5)

−1.14, 3.14

PEF (L/s)

7.15 (2.21)

90 (23)a

7.15 (1.90)

0. 00 (1.48)

−0.6, 0.6

FEF25–75 (L/s)

3.86 (1.09)

115 (28)a

3.96 (0.96)

0.10 (0.56)

−0.14, 0.34

The spirometry values are a percentage of normal values predicted for men and women of same height and age [32].

FEV1 = forced expiratory volume in 1 s, FVC = forced vital capacity, PEF = peak expiratory flow, FEF25–75 = forced expiratory flow over the middle half of expiration, SVC = slow vital capacity.

aPredicted value significantly different than measured value presurgery (p < 0.05)

bThe change between pre and postsurgery is significant (p < 0.05)

Alveolar-membrane diffusing capacity and DLCO was impaired presurgery (Table 5). Alveolar-membrane diffusing capacity improved by ∼10% postsurgery, such that the predicted values were no longer different compared to the measured values. The DLNO/DLCO ratio and VA both increased by ∼5%, and Vc/DmCO decreased by ∼10% (p < 0.05).
Table 5

Changes in diffusing capacity and its components with weight loss

n = 21 (women = 14; men = 7)

Presurgery

Percent predicted presurgery

Postsurgery

Change

95% confidence interval for the mean change

Mean (SD)

Mean (SD)

Mean (SD)

Mean (SD)

Diffusing capacity parameters

DLNO (mL min−1 mm Hg−1)

128.0 (28.3)

94 (14)b

139.3 (33.9)

11.3 (18.3)a

3.5, 19.1

DLCO (mL min−1 mm Hg−1)

25.8 (6.3)

97 (16)

26.5 (6.6)

0.8 (3.8)

−0.8, 2.4

DLNO/DLCO

5.00 (0.37)

--

5.26 (0.38)

0.26 (0.32)a

0.12, 0.40

VA (BTPS) (L)

5.08 (1.03)

97 (12)

5.32 (1.11)

0.24 (0.44)a

0.05, 0.43

DLNO/VA

25.2 (3.0)

26.1 (3.0)

0.9 (2.1)

0, 1.8

DLCO/VA

5.1 (0.6)

5.0 (0.5)

−0.1 (0.5)

−0.3, 0.1

Vc/VA (mL/L)

13.5 (2.0)

112 (16)b

12.9 (1.2)

−0.6 (1.6)

−1.3, 0.1

DmCO (mL min−1 mm Hg−1)

53 (12)

88 (23)b

58 (14)

5 (8)a

2, 8

Vc (mL)

75 (14)

98 (18)

68 (14)

1 (11)

−4, 6

Vc/DmCO

1.29 (0.14)

104 (11)

1.21 (0.14)

−0.09 (0.13)a

−0.15, −0.03

Predicted DLCO, DLNO, and Vc are from Zavorsky [33] and are based on height, gender, and age (mL/min/mmHg). Predicted alveolar-membrane diffusing capacity (DmCO) values are from Zanen [21] based on height, gender and age and converted to mL/min/mmHg. For predicted Vc/DmCO, the predicted Vc from Zavorsky [33] was divided by the predicted DmCO from Zavorsky [33]. Alveolar volume is predicted from Frans [34] based on height, gender, and age. The 95% confidence interval was calculated as the mean difference  ± 1.96 × standard error of the mean. The change in DLNO/VA was nearly significant (p = 0.07).

DLNO = pulmonary diffusing capacity for nitric oxide, DLCO = pulmonary diffusing capacity for carbon monoxide, VA = alveolar volume, DmCO = alveolar-membrane diffusing capacity for carbon monoxide which is calculated as DLNO/2.42, Vc = pulmonary capillary blood volume.

aThe change between pre and post surgery is significant (p < 0.05)

bPredicted value significantly different than measured value pre-surgery (p < 0.05)

Multiple linear regression was performed to discover the factors that were associated with the change in DLNO from bariatric surgery. The following predictors were used in the regression: gender and changes in AaDO2, %SaO2, waist circumference, hip circumference, VA, FVC, and weight. Stepwise regression revealed that the strongest predictors to the improvement in DLNO were the change in alveolar volume and the waist-to-hip ratio (Adjusted r2 = 0.76). Specifically, the improvement in alveolar volume and the reduction in the waist-to-hip ratio increased DLNO. The formula is the following: ΔDLNO (mL/min/mm/Hg) = 37.1 (Δ alveolar volume in liters) − 117.4 (Δ waist-to-hip ratio)−2.3 (adjusted r2 = 0.76, standard error of the estimate (SEE) = 9.2; F ratio = 30.1; Durbin–Watson statistic = 1.7; Mallows’ prediction criterion = 3.2; p < 0.001), where Δ represents the change, and the change is the postsurgical value subtracted by the presurgical value. Pearson product moment corrections demonstrated no relationship between the change in DLNO and the change in either waist (r = −0.071, p = 0.76) or hip circumference (r = 0.16, p = 0.50).

Discussion

In agreement with our hypothesis, we have found that alveolar-membrane diffusing capacity, as reflected by DLNO, improved with weight loss, and a majority of the improvement was associated with the improvement in alveolar volume and reduction in the waist-to-hip ratio. The novelty of this study comes from the use of DLNO to examine alveolar-membrane conductance and the association of the changes in waist-to-hip ratio to DLNO in the morbidly obese.

The improvement in alveolar volume from surgical weight loss was likely associated with a decrease in both the mechanical constraint and intra-abdominal pressure from the reduced abdominal fat mass placed on the thoracic cage. Therefore, because the relationship between the change in waist-to-hip ratio and the change in waist circumference was strong (r2 = 0.69; p < 0.01) and the relationship between the change in the waist-to-hip ratio and the change in hip circumference was not (r2 = 0.1, p = 0.18), the reduced waist-to-hip ratio comes from mainly the reduction in waist circumference. The decrease in waist circumference from surgical weight loss should reduce intra-abdominal pressure [24]. As a result, the diaphragm would be lowered, intrathoracic pressure would be reduced, and the alveolar volume would be increased.

As the waist-to-hip ratio is associated with pulmonary gas exchange [2], we also thought that the improvement in gas exchange from the reduction in the waist-to-hip ratio would contribute to the improvement in DLNO. However, stepwise regression did not show that the small improvement in gas exchange was associated with the improvement in DLNO. It could be that the presurgical values of %SaO2 were already normal and the increased AaDO2 presurgery was not large enough such that, when weight loss ensued, the changes were not significant to affect DLNO.

It may be surprising that a predictor of the multiple linear regression included the waist-to-hip ratio but not the waist circumference. Because increases in intra-abdominal pressure correlate with the waist circumference and poorly with the waist-to-hip ratio [24] (due to the dilutional effect of the hip circumference in obese patients who have both central and peripheral obesity), it may be assumed then that the waist circumference would be a better predictor of the improvement in DLNO compared to the waist-to-hip ratio. However, this was not this case. Previous work illustrated that pulmonary gas exchange is more associated to the waist-to-hip ratio than the waist circumference alone [2] and, thus, we felt that this would also be true for DLNO. Clearly, the debate as to whether the waist-to-hip ratio or waist circumference is a better or more physiologically appropriate predictor of gas exchange and diffusing capacity is not resolved.

In contrast to our data, it has been shown that DLCO is 8% higher in patients with morbid obesity [25, 26]. However, we did not see a higher DLCO in our subjects prior to surgery. In fact, the DLCO was similar to predicted values. The increase in DLNO from surgery could not be due to increases in cardiac output, as both heart rate and oxygen consumption (L/min) at rest were significantly lower postsurgery (p < 0.01).

Indeed, whereas the more recent data show that DLCO is elevated in those with morbid obesity, there is still debate. Other studies show unaltered or slightly lower DLCOs compared to those predicted [2730], and like ours, the effect of about the same drop in BMI on DLCO was insignificant [29]. What is novel in this study is that we have taken an extra step and have examined the changes in the components of diffusing capacity, namely pulmonary capillary blood volume and alveolar-membrane diffusing capacity as reflected by DLNO, and showed that, despite a normal DLCO, the alveolar-membrane conductance was lower than predicted. This suggests that the sensitivity to detect change may be better with DLNO compared with DLCO [20].

There is only one other study that we know that has also looked at pulmonary capillary blood volume and alveolar-membrane conductance in the morbidly obese, and these authors also showed a larger than predicted pulmonary capillary blood volume and a lower than predicted alveolar-membrane conductance [30]. However, that study used the traditional two-step Roughton–Forster method to obtain these parameters, and that method has some disadvantages, which are described elsewhere [4, 5]. Another problem noticed is that it was difficult to find out where the predicted norms were obtained in those studies [2730]. This is a major issue, as it is difficult to contrast and compare diffusion data between studies. In this study, we have made sure to report the measured values and where the predicted norms come from for future comparison.

One limitation of this study is that one-third of the subjects actually showed a decrease in DLNO and DLCO postsurgery. In fact, the change in DLNO was highly related with the change in DLCO (r2 = 0.85), even when normalized for alveolar volume (r2 = 0.69). So when DLNO decreased postsurgery, DLCO was also likely to be decreased postsurgery. One reason for the worsening of diffusing capacity postsurgery could be that the postoperative clinical status could have affected the results. Specifically, the postoperative complications could have affected lung function. However, the classification of surgical complications, which rates the level of intervention needed postsurgery based on the Clavien system [31], was not related to the decrease in diffusing capacity. This means that those with a decrease in diffusing capacity postsurgery were as likely to have worse complications as those whose diffusing capacity improved postsurgery.

In conclusion, alveolar-membrane conductance, as reflected by DLNO, normalizes within 10 weeks after bariatric surgery in those who are morbidly obese. This is likely due to the increase in alveolar volume from the reduction in the waist-to-hip ratio.

Acknowledgments

This data was initially presented as a poster at the 2007 Canadian Anesthesiologist’s Society (CAS) annual meeting in Calgary, Alberta. It was awarded as one of the top 50 abstracts of the conference. The authors would like to thank the CAS for partially funding this project. The authors would also like to thank the volunteers for being a part of this research study, and the Anesthesiologists’ from the Department of Anesthesia at the Montreal General Hospital who inserted the arterial lines in the subjects.

Copyright information

© Springer Science + Business Media B.V. 2007