Oxygen desaturation during maximal stair-climbing test and postoperative complications after major lung resections Free

Abstract

Objective: Non-univocal conclusions have been published regarding the definition of oxygen desaturation in relation to postoperative outcome. We aimed to verify whether oxygen desaturation during a maximal stair-climbing test was associated with postoperative cardiopulmonary complications and to assess which definition of oxygen desaturation (oxygen saturation <90% or desaturation >4% with respect to rest level) discriminated better between complicated and uncomplicated patients. Methods: Five hundred and thirty-six patients performing a maximal stair-climbing test prior to major lung resection were analyzed. All patients performed the test on room air. Patients with and without cardiopulmonary complications were compared in terms of several preoperative and operative characteristics by univariate analysis, including the presence of oxygen desaturation at peak exercise (saturation <90% or desaturation >4%). Logistic regression analysis was then performed and validated by bootstrap procedure to identify predictors of complications and to see whether the exercise oxygen desaturation retained its significancy after multivariable adjustment. Results: Twenty-seven patients had an exercise oxygen saturation below 90%, but this parameter was not significantly associated with complications. Seventy-five patients experienced an exercise desaturation greater than 4%, which was a significant result associated with postoperative complications at univariate analysis (p = 0.008) (36% complication rate). After adjusting for age, ppoFEV1, ppoDLCO, type of operation, height reached at stair-climbing test and cardiac co-morbidity, a desaturation greater than 4% retained its significance at logistic regression and proved to be stable at bootstrap. Conclusions: A stair-climbing test is an intense constant workload exercise, challenging a large amount of muscle mass, and appears particularly appropriate to elicit oxygen desaturation, which in turn may be a reliable marker of deficits in the oxygen transport system. A desaturation >4% appears a better cut-off definition than a saturation level <90% in predicting the occurrence of complications. The risk of complications was approximately two-fold higher in patients with oxygen desaturation >4% at peak exercise.

1 Introduction

Exercise tests are increasingly used in the preoperative evaluations of lung resection candidates [1,2]. Yet, the role of exercise oxygen desaturation (EOD) in risk-stratification has not been defined and non-univocal conclusions have been published regarding its definition and its association with early outcome after lung resection [3–5].

Due to the current demand for financial accountability for medical tests we elected to implement a low-technology inexpensive exercise test – stair-climbing test – to screen patients for operation. We have previously shown that a symptom-limited stair-climbing test was a valid predictor of postoperative complications, inasmuch as it is a stressful form of exercise that is capable of revealing deficits in the oxygen transport system [6].

In the present study we wanted to assess the role of EOD during stair-climbing test in predicting cardiopulmonary complications after major lung resection and define which is the most valuable definition of EOD (saturation level <90% or desaturation >4% with respect to rest value) for the purpose of risk-stratification.

2 Patients and methods

We analyzed 536 patients (110 females, 426 males) who performed a maximal stair-climbing test prior to major lung resection (440 lobectomies, 96 pneumonectomies) for lung cancer during a 6-year period (January 2001 through December 2006).

All consecutive patients submitted to major lung resections during the study period were deemed eligible for the study, provided they were able to perform a stair-climbing test. Previous thoracic procedures (18 patients) or cardiac operations (12 patients) were not considered exclusion criteria for the study.

This is an observational analysis performed on a prospectively compiled electronic database. Informed consent was obtained from all patients to use their data for clinical research. The database was approved by the local institutional review board.

All patients underwent a preoperative functional evaluation including: cardiologic evaluation, pulmonary function studies including measurement of carbon monoxide lung diffusion capacity, maximal stair-climbing tests and, in selected patients, cycle-ergospirometry with VO₂ peak assessment.

Criteria for inoperability were a predicted postoperative FEV1 (ppoFEV1) and predicted postoperative DLCO (ppoDLCO) less than 30% of predicted normal values associated with an insufficient exercise tolerance (height at stair-climbing test less than 12 m and VO₂ peak <10 ml/kg/min) [6].

During the study period another 70 patients were operated on without performing a preoperative stair-climbing test for medical contraindication (severe incapacitating skeletal, neurological, vascular, psychiatric diseases, cachexia) or refusal to perform the test.

As a rule, lung resections were performed through a muscle sparing lateral thoracotomy by qualified thoracic surgeons. Postoperative treatment was standardized and focused on early mobilization, chest physiotherapy and physical rehabilitation, thoracotomy pain control, antibiotic and anti-thrombotic prophylaxis. Postoperative chest pain was controlled by means of continuous intravenous analgesia (ketorolac and tramadol), which was titrated to keep the pain visual score below 4 (in a scale ranging from 0 to 10) during the first postoperative 48–72 h (pain score was assessed twice daily, during morning and afternoon rounds). This regimen was switched to oral administration of paracetamol after 48–72 h.

Postoperative morbidity and mortality were considered as those occurring within 30 days postoperatively or for a longer period if the patient was still in the hospital.

A number of preoperative and operative variables were tested for possible association with outcome (see  Appendix B for explanation of variables).

2.1 Stair-climbing test

All the patients in the study completed a symptom-limited stair-climbing test preoperatively.

Our hospital has 16 flights of stairs, each one constituted of 11 steps. Each step is 0.155 m high. The patients were asked to climb, at a pace of their own choice, the maximum number of steps and stop only for exhaustion, limiting dyspnea, leg fatigue or chest pain. All patients were accompanied by a physician who interacted with them and assessed their symptoms. The following parameters were recorded before and immediately after the exercise: arterial blood pressure, heart rate, and respiratory rate. During the exercise the patients’ pulse rate and capillary oxygen saturation were monitored by means of a portable pulse oxymeter with finger probe. All tests were performed on room air. The minimum value of oxygen saturation observed during the test was recorded and used to calculate the desaturation level from the rest oxygen saturation value. Two cut-off definitions of oxygen desaturation were tested: an oxygen saturation level below 90% and a desaturation greater than 4% (corresponding to the mean oxygen desaturation + 1SD in the present series).

3 Statistical analysis

Several preoperative and operative variables were tested for a possible association with postoperative cardiopulmonary morbidity. Variables were initially screened by univariate analysis. The univariate comparisons of outcomes were performed by means of the unpaired Student’s t-test for numerical variables with normal distribution and by means of the Mann–Whitney U-test for numerical variables without a normal distribution. The Shapiro–Wilk normality test was used to assess normal distribution. Categorical variables were compared by means of the Chi-square test or the Fisher’s exact test, as appropriate.

Variables with a p < 0.10 at univariate analysis were then used as independent variables in the stepwise logistic regression analysis. The presence or absence of one or more complications was used as dependent variable. All data were complete with the exception of DLCO, which was 95% complete. Missing data were imputed by averaging the non-missing values. To avoid multicollinearity, only one variable in a set of variables with a correlation coefficient greater than 0.5 was selected (by bootstrap procedure) and used in the regression model. A p < 0.1 was selected for retention of variables in the final model.

Logistic regression was then validated by bootstrap analysis with 1000 samples. In the bootstrap procedure, repeated samples of the same number of observations as the original database (536) were selected with replacement from the original set observations. For each sample, stepwise logistic regression was performed entering the variables with p < 0.1 at univariate analysis. The stability of the final model can be assessed by identifying the variables that enter most frequently in the repeated bootstrap models and comparing those variables with the variables in the final model. If the final stepwise model variables occur in a majority (>50%) of the bootstrap models, these are judged to be stable [7–9].

All the statistical tests were two-tailed and a significance level of 0.05 was accepted. The analysis was performed by using the STATA 8.2 (Stata Corp., College Station, TX) statistical software.

4 Results

Table 1 shows the characteristics of the patients in the study.

Cardiopulmonary and mortality rates were 24% (128 cases) and 3.7% (20 cases), respectively. Complication and mortality rates were 22% and 2.2% after lobectomy, and 32% and 10% after pneumonectomy, respectively.

Complications in order of frequency included the following: arrhythmia (46 cases), pneumonia (34 cases), respiratory failure (17 cases), atelectasis (12 cases), cardiac failure (11 cases), myocardial ischemia (nine cases), pulmonary edema (three cases), stroke (two cases), acute renal insufficiency (two cases), pulmonary embolism (one case).

Mean oxygen desaturation during exercise was 1.6% (standard deviation 2.8). Ninety-four patients (18%) improved their saturation level during exercise, whereas 328 (61%) had an oxygen desaturation of some extent.

Correlation coefficients between changes in oxygen saturation from rest to peak exercise and age, FEV1, DLCO and height at stair-climbing test were all below 0.2, indicating poor correlation.

Twenty-seven patients (5%) reached an exercise saturation <90%, whereas 75 patients (14%) had an EOD > 4%.

The following variable results were associated with cardiopulmonary morbidity at univariate analysis (Table 2 ): age (p < 0.0001), male gender (p = 0.002), ppoFEV1 (p = 0.0001), ppoDLCO (p = 0.03), pneumonectomy (p = 0.03), presence of cardiac co-morbidity (p < 0.0001), height reached at stair-climbing test (p = 0.001), and exercise oxygen desaturation >4% (p = 0.008).

An exercise saturation level below 90% was not found to be significantly associated with postoperative complications.

Stepwise logistic regression analysis showed that significant and reliable factors independently associated with postoperative cardiopulmonary morbidity were age (p = 0.003), ppoFEV1 (p = 0.004), presence of cardiac co-morbidity (p = 0.002), height at stair-climbing test (p = 0.045), and EOD > 4% (p = 0.05). All these factors proved to be stable at bootstrap analysis (Table 3 ).

The complication rates in patients with and without EOD > 4% were 36% and 22%, respectively (p = 0.008).

In this series, 37 patients climbed less than 12 m in the stair-climbing test. Of these, seven patients had EOD > 4%. The morbidity rate in this high-risk group was 71% (five cases). Conversely, of the 499 patients climbing higher than 12 m, 68 had EOD > 4%, and among these 22 had complications (32%). Of those patients without EOD, those climbing less than 12 m had a morbidity rate of 37% versus 20% of those climbing higher than 12 m.

One hundred and thirty-four patients (25%) had moderate to severe COPD, according to GOLD criteria (FEV1 < 80% and FEV1/FVC ratio < 0.7). In these patients, 31 experienced EOD > 4% with 11 complicated (35%). Of those COPD patients without EOD, 29 had complications (28%).

Two hundred and fifty-eight patients (48%) had a concomitant cardiac disease. In this group, 35 patients experienced EOD > 4%, with a morbidity rate of 51% (18 cases). Patients with cardiac co-morbidity but without EOD had a morbidity rate of 29% (65 cases).

Nine of the 75 patients with EOD > 4% (12%) needed a postoperative emergency admission to the intensive care unit. This rate was three-fold higher than in patients without EOD (p = 0.003).

Six of the 20 patients who died had EOD > 4%. The mortality rate in patients with EOD > 4% was 8% (nine of 75) versus 3% in those without EOD (p = 0.04).

Among patients with EOD > 4%, those who climbed less than 12 m had a mortality rate of 29% (two of seven cases), versus a mortality rate of 6% in those who climbed higher than 12 m (four of 68 cases). In patients without EOD, the mortality rates were 10% (three of 30 cases) and 2.6% (11 of 431 cases) in those climbing lower or higher than 12 m, respectively.

In the five patients with cardiac co-morbidity, who climbed less than 12 m and experienced EOD > 4%, the morbidity and mortality rates were 80% (four cases) and 40% (two cases), respectively.

5 Discussion

Although exercise oximetry has been proposed as a useful tool in the preoperative functional evaluation of the lung resection candidates [1,2], non-univocal results have been reported on the role of exercise oxygen desaturation in preoperative risk-stratification before lung resection [3–5]. Although some authors reported that EOD was a good predictor of postoperative complications [3,4], others did not find similar results [5,10]. Differences in variable definition (saturation level <90% or desaturation >4% with respect to rest value) and in the exercise tests used may account for these discrepancies.

The objectives of our study were to verify the most valuable definition of EOD (either a saturation level <90% or desaturation greater than 4% with respect to rest value) in relationship to risk-stratification and whether this parameter was a reliable predictor of postoperative cardiopulmonary morbidity.

We chose the stair-climbing test as a form of maximal exercise test, inasmuch as it is safe, economical, and widely applicable. This test is familiar to the patients, needs few personnel and equipment, and was shown to be useful in predicting cardiopulmonary complications after lung resection [6]. Moreover, it has been reported that stair-climbing is a more stressful exercise than cycle or treadmill [11–13] and for this reason it is a valid tool for detecting abnormalities in the oxygen transport system.

Oxygen saturation during the test was monitored by means of a portable pulse oxymeter by using a finger probe. This technique has been shown to accurately estimate changes in arterial saturation between rest and exercise [14].

We chose two definitions of EOD: a saturation level below 90%, as most commonly accepted in the literature [3–5,10,15], and a decrease in oxygen saturation greater than 4%, corresponding in our case to the mean oxygen desaturation plus one standard deviation. We arbitrarily chose one SD to estimate the EOD cut-off, inasmuch as it is the most commonly used measure of variability of a set of numbers in the same units of measurement as the original observations. Our cut-off was similar to the one reported by other authors and selected with other distributional techniques [3,4].

We found that, whereas a saturation level below 90% during exercise was not associated with postoperative morbidity, a desaturation greater than 4% was a reliable predictor of cardiopulmonary complications even after adjusting for other preoperative and operative factors by logistic regression analysis.

EOD > 4% proved to be an important marker for adverse outcome particularly in those patients with an underlying cardiac disease. In these patients, the morbidity rate was as high as 51%.

We also found that 71% of those patients with a poor performance at the stair-climbing test, corresponding to a height lower than 12 m [6], and with EOD > 4% developed postoperative complications. In this group the mortality rate was almost five-fold higher than in patients with EOD > 4% but climbing more than 12 m and more than 10-fold higher than in those without EOD and climbing more than 12 m.

In patients with cardiac co-morbidity, who climbed less than 12 m and experienced EOD > 4%, the morbidity and mortality rates were 80% and 40%, respectively. These very high-risk groups of patients need to be identified before operation and their perioperative management optimized.

It appears clear that five or more points of desaturation indicate severe abnormalities in the oxygen transport system, irrespective of the actual level of oxygen saturation (above or below 90%).

Oxygen desaturation during exercise may in fact be explained by one or a combination of the following factors: alveolar ventilation does not rise relatively as much as VO₂; the cardiac output response to exercise may be subnormal (for coexisting ischemic heart disease or pulmonary hypertension), such that the mixed venous oxygen tension is very low, and, in the presence of ventilation/perfusion inequality, this will depress the arterial oxygen tension; exercise could lead to ventilation/perfusion mismatching (most probably due to the temporary accumulation of interstitial fluid in the lungs for an increased hydrostatic vascular pressure in both pulmonary artery and vein) [16].

Based on the results of this study, we think that all patients experiencing an exercise oxygen desaturation greater than 4% should undergo a formal cardiopulmonary exercise test (CPET) with VO₂ peak measurement before the operation in order to define the mechanisms impairing their oxygen transport system (cardiac, pulmonary or pulmonary vascular).

This study is to our knowledge the largest prospective series investigating the importance of exercise oxygen desaturation on the outcome of a homogeneous group of patients submitted to major lung resection for lung cancer. However, it may have potential limitations. First, this was an observational analysis and although data have been prospectively collected, patients were not randomly allocated to surgery. Inadvertent inherent selection biases are always to be considered in this type of analysis [17]. Second, the reproducibility of the results generated in this study in other settings or with other types of exercise tests needs to be verified. Third, mortality was not separately assessed by regression analysis due to the relative rarity of events. Although EOD appeared to be associated with mortality at univariate analysis, further investigations with larger sample sizes allowing a reliable multivariate regression are needed. Finally, exercise tests cannot be performed by all lung resection candidates. For these patients, exercise parameters, among which EOD, are not available for risk-stratification and alternative methods of risk-assessment should be researched [18].

In conclusion, we found that an oxygen desaturation greater than 4% (but not a saturation level below 90%) at maximal stair-climbing test was reliably associated with postoperative cardiopulmonary complications. In our view, all lung resection candidates should perform at least a preoperative low-technology exercise test with oxygen saturation monitoring. In this regard, we think a stair-climbing test is a cost-effective firstline screening test challenging a large amount of muscle mass, and for this reason, particularly appropriate to elicit oxygen desaturation. This in turn may be a reliable marker of deficits in the oxygen transport system. Those patients in whom EOD is observed should be further assessed with a formal CPET for a more precise evaluation of their cardiopulmonary system in order to optimize their perioperative management. Nevertheless, other studies are needed to confirm our results in other settings and with other types of exercise tests and to assess the association of oxygen desaturation with long-term outcomes such as residual quality of life and survival.

Preoperative and operative variables: The following variables were initially screened for a possible association with postoperative morbidity: age, gender, body mass index (BMI, kg/m²), smoking history (pack-years), COPD status (GOLD criteria), forced expiratory volume in 1 s (FEV1), carbon monoxide lung diffusion capacity (DLCO), predicted postoperative FEV1 (ppoFEV1), predicted postoperative DLCO (ppoDLCO), cardiac co-morbidity, type of operation (lobectomy vs pneumonectomy), neoadjuvant chemotherapy; stage pT (pT1 vs pT > 1), stage pN (pN negative vs pN positive), height reached at maximal stair-climbing test (m); exercise oxygen saturation level below 90%; exercise oxygen desaturation greater than the mean oxygen desaturation in the entire population (1.6%) + 1 standard deviation (2.8) (corresponding to a value of or greater than 5%). Pulmonary function tests were performed according to the American Thoracic Society criteria. Results of spirometry were collected after bronchodilator administration. DLCO measurement was performed by the single breath method.

FEV1, ppoFEV1, DLCO and ppoDLCO were expressed as percentages of predicted for age, sex and height. PpoFEV1 and ppoDLCO were calculated by taking into account the number of functioning segments removed during operation. Quantitative lung perfusion scan was also performed in all pneumonectomy candidates according to published guidelines [1].

For the purpose of the present study and in accordance with previous investigations [19,20], a concomitant cardiac disease was defined as follows: previous cardiac surgery, previous myocardial infarction, history of coronary artery disease, current treatment for arrhythmia, cardiac failure or hypertension.

Outcome variables: For the purpose of this study, according to previous studies [20–22][20-22] and to the EACTS/ESTS thoracic surgery database [23], the following complications were included: respiratory failure requiring mechanical ventilation for more than 48 h; pneumonia (chest X rays infiltrates, increased white blood cell count, fever); atelectasis requiring bronchoscopy; adult respiratory distress syndrome (ARDS); pulmonary edema; pulmonary embolism; myocardial infarction (suggestive electrocardiogram findings and increased myocardial enzymes); hemodynamically unstable arrhythmia requiring medical treatment; cardiac failure (suggestive chest X rays, physical examination and symptoms); acute renal failure (change in serum creatinine greater than 2 mg/dl compared to preoperative values); stroke.

References

Author notes