Risk Stratification (Thoracic Anesthesia)

Central to the preoperative assessment of thoracic anesthesia patients are two concepts: first is the idea that it is the thoracic surgeon, and not the anesthesiologist, who ultimately determines resectability (although the anesthesiologist may function as an important safety check. Furthermore, the anesthetic consultant may be asked to opine in certain intraoperative situations, ex. requirement for a larger resection). Second is the idea that there are no “elective” operations – for any procedure, the risks of delaying the operation (metastasis, growth at resection site, etc.) must be weighed heavily

Traditionally, quantitative preoperative assessment has been thought of as the “three legged stool” which includes 1) respiratory mechanics [FEV1, although the patient’s preoperative quality of life is probably the best assessment of respiratory function] 2) lung parenchymal function [DLCO] and 3) cardiopulmonary reserve [VO2 Max]. Also can be thought of as getting oxygen, exchanging oxygen, and delivering oxygen

Other important considerations are to 1) determine maximal extent of resection 2) identify who will need an ICU bed [i.e. post-operative ventilatory support] and 3) determine the effectiveness of bronchodilator therapy [i.e. should you use it post-operatively] and 4) identify difficult airways

Respiratory Mechanics (getting oxygen)

Always inquire about quality of life, as it may be the best indicator of respiratory function. That said, certain tests are always in order


Of all the spirometric exams which correlate with pulmonary complications (FVC, MVV, RV/TLC, FEV1%, ppoFEV1%), ppoFEV1 is the most predictive of pulmonary complications [British Thoracic Society. Thorax 56: 98, 2001]

ppoFEV1% = preoperative FEV1% x (1 – %functional lung tissue removed/100)

Generally, a ppoFEV1% > 40% is low-risk for respiratory complications, and ppoFEV1% < 30% is high risk [Nakahara K et al. Ann Thor Surg 46: 549, 1988, 156 patients]. Note that because of compensatory hyperinflation of the residual lung, eventual FEV1 will exceed ppo FEV1 by 250 mL in lobectomies and 500 mL in pneumonectomies [Zehier B et al. Chest 105: 753, 1995]

By spirometry, a vital capacity (maximal ins./exp.) of < 50% of predicted or < 2L is predictive of increased risk [Gass GD and Olsen GN. Chest 89: 127, 1986] – according to Barash, “abnormal” vital capacity yields a 10% risk of mortality. FEV1 may be a better indicator than vital capacity – if FEV1 is > 2L, mortality is ~ 10%, whereas for FEV1 < 1L, mortality may increase to > 20% [Lockwood P. Respiration 30: 529, 1973]. Maximum voluntary ventilation < 50%, and RV/TLC > 50% [Mittman C. Am Rev Respir Dis 84: 197, 1961] may also be predictive

Keep in mind, however, that some of the above data are based on the pre-thorascopy and pre-epidural eras, and may no longer be as applicable. For instance, classically FEV1 < 800 cc used to be an absolute contraindication to thoracic surgery, but no longer is. Note also that in order to cough effectively, VC ≥ 3 x TV.

Flow-Volume Loops

Graphical output of a spirometric measurement. Actual volumes are on the x-axis, flow rates are on the y-axis. Note that the x-axis is numbered in reverse (0 is to the right). Thus, patients with obstructive disease, whose lungs exist at larger volumes, will be shifted to the left.

The F-V loop begins at FRC (furthest to the right). Inspiration is represented by the fraction of the F-V loop below the x-axis, exhalation is represented by the fraction above it. The width of a flow-volume loop is equal to vital capacity (which is why patients with restrictive lung disease have very narrow F-V loops)

Expiration in the flow-volume loop is made of two components, an effort-dependent component (accelerating flow, dependent on larger airways [lung volumes are large]) and an effort-independent component (decelerating flow, dependent on smaller airways as the lung volumes are smaller)

Lung Parenchymal Function (exchanging oxygen)

Clasically, thresholds of PaO2 < 60 mm Hg and PaCO2 > 45 mm Hg were used to assess (and rule out) thoracic surgery patients. The combination of improved surgical technique and the DLCO study have obviated these cutoff values, although they should still serve as red flags

Diffusion Capacity for CO (DLCO)

Predicted post-operative DLCO is the single strongest predictor of complications and mortality after lung resection [Barash], although it is important to note that DLCO is NOT predictive of long term survival, only perioperative mortality [Wang J et al. J Thorac Cardiovasc Surg 17: 5811, 1999]. Interestingly, ppoDLCO and ppoFEV1 are poorly correlated, and thus should be assessed independently [Ferguson MK et al. J Thor Cardiovas Surg 109: 275, 1995]

ppoDLCO = preoperative DLCO x (1 – %functional lung tissue removed/100)

A ppoDLCO < 40% is correlated with increased cardiopulmonary complications. In the National Emphysema Treatment Trial, which included 1218 patients, “high risk” patients (FEV1 < 20% and either ppoDLCO < 20% or homogenous emphysema) showed a 35% risk of death at one year in those who had surgery (as compared to 10% in those who did not) [National Emphysema Treatment Trial Research Group. NEJM 348: 2059, 2003]

Cardiopulmonary Reserve (delivering oxygen)

VO2 max

VO2 max can also be used to predict postoperative complications, although, somewhat counter intuitively (i.e. one would expect a patient with a high VO2 max to require high amounts of oxygen), a high VO2 max predicts good outcomes.

Patients with a VO2 max < 10 mL/kg/min are at very high risk [Bechard D et al. Ann Thorac Surg 44: 344, 1987; Bollinger CT et al. Chest 108: 341, 1995], and 15 mL/kg/min may be a reasonable cutoff [Miller’s Anesthesia, 7th ed. 2009. p 1821]

Stair Climbing

Because calculation of VO2 max is expensive, stair climbing has been proposed as an alternative. It is commonly cited that the ability to climb five flights of stairs without stopping (20 x 6” steps) is equivalent to a VO2 max of 15 mL/kg/min, and two flights correspond to 12 mL/kg/min, however this data is based on Pollock’s data which showed a reasonable correlation between steps climbed and (r = 0.7), however there was significant scatter at each point [Pollock M et al. Chest 104: 1378, 1993] – indeed, VO2 max ranged from 9 to 19 mL/kg/min at 50 steps, and from 15 to 35 mL/kg/min at 90 steps. Pollock himself concludes that the 1-2 flight criteria is inadequate to evaluate cardiopulmonary reserve. An evaluation of 54 patients undergoing thoracic surgery, found that the ability to climb three flights was most predictive of post-operative complications [Olsen GN et al. Chest 99: 587, 1991]

6 Minute Walk Test

Has been shown to correlate with VO2 max. Less than 2000 feet (610 meters) suggests as VO2 max < 15 mL/kg/min [Miller’s Anesthesia, 7th ed. 2009. p 1821], and also correlates with a fall in SpO2 [Miller’s Anesthesia, 7th ed. 2009. p 1821]. This is important, because drops in SpO2 > 4% during exercise are associated with increased morbidity [Ninan M et al. Ann Thorac Surg 64: 328, 1997]

Other Measurements / Calculations

Spontaneous respiratory rate is the most sensitive clinical index of lung compliance, as patients with low compliance will compensate by taking rapid, shallow breaths

Ventilation-Perfusion Testing

Previous guidelines have not recommended V-Q scanning, however these guidelines were based on relatively outdated techniques – Win et al. studied modern V-Q scanning in 61 patients, and found a significant correlation between both scintigraphic and simple segment counting-based predicted post-operative lung function and actual post-operative pulmonary function [Win T et al. Ann Thor Surg 78: 1215, 2004]. Importantly, Win et al. did not feel that quantitative V-Q scanning was necessary, as the counting method seemed to suffice

The counting method is as follows:

ppoFEV1 = preop FEV1 × (1 − segments of lung to be resected/19)

as described by Bolliger and colleagues [Bolliger CT et al. Respiration 69: 482, 2002], with 19 being the total number of functional lung segments. Keep in mind, however, that V-Q scanning is always conducted at rest, whereas ultimate lung function is dependent on maximal activity, thus there is always some inherent error in V-Q predictions

Split-Lung Function Tests

The rationale of regional lung function tests is that global pulmonary function is irrelevant post-resection. Regional perfusion tests (using IV 133Xe) allow one to examine the relative perfusion of each lung. Insoluble, inhaled radioactive-labeled gasses (xenon, 99m-technetium) can be similarly used to conduct regional ventilation tests. The most “functional” split-lung function test is the bronchial balloon occlusion test, in which the segment to be resected is blocked and traditional pulmonary function tests are then repeated (and compared to the original)

PA Balloon Occlusion Test

In the past a PA catheter was advanced into the lung section of interest, and used to occlude blood flow – if mean PAP increased to more than 40 mm Hg, or if PaO2 < 60 mm Hg or PaCO2 > 45 mm Hg, postoperative respiratory failure or cor pulmonale was likely. A study of 20 patients, however, showed that PA clamping led to an increased PA pressure in only 50% of patients, but 70% had a decreased RVEF [Lewis JW et al. J Thorac Cardiovasc Surg 108: 169, 1994]. Thus, PA occlusion/clamping tests may miss a significant number of patients who suffer adverse RV consequences when the pulmonary vasculature is disrupted, and because of the usefulness of split function studies and other tests, are rarely performed anymore

Past Medical History


There is no age cutoff for thoracic surgery – in fact, Osaki et al.’s series of 30 octogenarians with lung cancer (age range 80-92) showed an operative mortality rate of 3%, and a five year survival rate of 32%, although perioperative complication rates were higher than in reported values for younger patients [Osaki et al. Ann Thor Surg 57: 188, 1994]. That said, the mortality rate of elderly patients undergoing pneumonectomies (especially R pneumonectomy) is high, but whether or not pneumonectomies should be prohibitive in the elderly is controversial – in one group of 9 patients > 75 years, none survived at 5 years following a pneumonectomy, however in another study of 35 patients, a 17.5% 5-year survival rate was achieved [Spaggiari L and Scanagatta P. Curr Opin Oncol 19: 84, 2007]

Cardiovascular Disease

Thoracotomy is considered an “intermediate risk” procedure, however cardiovascular complications are the second most common cause of morbidity and mortality after thoracic surgery, with a 5% incidence of perioperative ischemia (peaks at 2-3 days). That said, routine invasive cardiovascular testing is not cost-effective, nor has it been shown to be helpful

As many as 50% of COPD patients will have RV dysfunction [Miller’s Anesthesia, 7th ed. 2009. p 1825], mostly due to chronic hypoxemia. Any patient with a ppoFEV1 < 40% should probably have a TTE to assess the RV, as elevated RV pressures place one in a particularly high surgical risk category

Pulmonary Disease

Patients with moderate or severe COPD may have bullae, which, as long as they take up less than 50% of a hemithorax, are often clinically silent. The risk of rupture, however, is significant, and thus if identified, positive pressure ventilation should be used with caution (airway pressures should be as low as possible, chest tube ready and lung isolation equipment immediately available)

Severely flow-limited patients are at risk for complete hemodynamic collapse following the initiation of positive pressure ventilation, even by bag-mask apparatus [Myles PE et al. Br J Anaesth 74: 340, 1995] – this occurs because they simply cannot exhale all of the forced inhalate, and the only effective treatment may be cessation of ventilation [Ben-David B et al. Anesth Analg 92: 690, 2001; FREE Full-text at Anesthesia & Analgesia]

COPD patients are also at risk for auto-PEEP, especially while under mechanical ventilation. Unfortunately, most standard anesthesia ventilators cannot detect auto-PEEP (this requires end-expiratory flow interruption). Auto-PEEP is proportional to tidal volume and inversely proportional to expiratory time, thus, if it is suspected, consider lowering TV and increasing the expiratory time

Patients with carcinoid tumors need to be watched carefully, as these can lead to refractory hypotension (may require octreotide or somatostain [Vaughan DJ and Brunner MD. In Anesthesiol Clin 35: 129, 1997]) or coronary artery spasm [Metha AC et al. Chest 115: 598, 1999]

Tumor/Cancer Assessment

Four aspects of disease must be assessed (the “Four M’s”) – mass effects (SVC, Pancoast, obstructive pneumonia, laryngeal nerve paralysis, phrenic paresis), metabolic effects (hypercalcemia, hyponatremai, cushing’s, lambert-easton), metastases (ex. adrenal), and medications (bleomycin [avoid high FiO2], cisplatin [avoid NSAIDs])


Assessment of patient anatomy is important in order to anticipate a difficult endotracheal, or endobronchial intubation. In addition to the physical exam, chest X-rays, CT scans, and bronchoscopy reports can all be of use.