| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2007;83:425-432
© 2007 The Society of Thoracic Surgeons
a Division of Thoracic Surgery, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts
b Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts
c VA Medical Center, Boston, and Harvard Medical School, Boston, Massachusetts
d University of Colorado Health Outcomes Program, Denver, Colorado
Accepted for publication September 25, 2006.
* Address correspondence to Dr Linden, Division of Thoracic Surgery, Brigham and Women’s Hospital, 75 Francis St, Boston, MA 02115 (Email: plinden{at}partners.org).
Presented at the Forty-second Annual Meeting of The Society of Thoracic Surgeons, Chicago, IL, Jan 30–Feb 1, 2006.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
METHODS: The Veterans Affairs National Surgical Quality Improvement Program Database was queried for all patients who underwent lobectomy, followed by an additional lung resection, between 1994 and 2002. Preoperative variables, intraoperative variables, and complications were analyzed. Pulmonary function data were not collected.
RESULTS: Excluding 17 patients who underwent repeat resection for complications of lobectomy, 186 patients underwent 191 repeat resections. The 30-day mortality was 11%; the complication rate was 19%. Mortality for pneumonectomy was 34%, lobectomy, 7%; segmentectomy, 0%; and wedge resection, 6%. The most frequent complications were pneumonia (9%), reintubation (8%), ventilator dependence (6%), cardiac arrest (3%), dysrhythmia (3%), and sepsis (3%). Multivariate analysis revealed that operative time exceeding 2 hours, preoperative dyspnea at rest or with minimal exertion, and white blood cell count of more than 10,000/mm3 were predictors of complication. Presence of a contaminated/infected case, pneumonectomy, and intraoperative transfusion were predictors of death. Age, complications from prior lobectomy, time interval between lobectomy and repeat resection, smoking history, other comorbidities, and preoperative laboratory values were not independent predictors.
CONCLUSIONS: Repeat lung resection after lobectomy carries an 11% overall mortality predicted by the presence of a contaminated/infected case, need for intraoperative transfusion, and pneumonectomy versus a lesser resection.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Prior retrospective studies suggest a higher risk of death (range, 5% to 9%) and complication after repeat resection compared with the initial resection [4–7]. Lung resection after lobectomy has not been prospectively studied, and it is not known if the risk factors for repeat lung resection are the same as those for initial resection. Our aim was to use a prospective database to quantitate the risk of complications resulting from repeat lung resection in patients who had previously undergone lobectomy.
The National Surgical Quality Improvement Program (NSQIP) is a national, multi-institutional Veterans Affairs (VA) Project designed to prospectively collect data and record perioperative outcome from a variety of surgical procedures. We examined 208 lung resections in 203 patients who had previously undergone lobectomy. Preoperative and intraoperative variables were assessed for their effect on 30-day morbidity and mortality. We evaluated the effect of these variables and the influence of prior lobectomy complications on the morbidity and mortality of repeat lung resection.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Population
The NSQIP database was queried for all patients who underwent pulmonary lobectomy, followed at a later date by an additional lung resection, between 1994 and 2002, and 253 patients were found. Patients were excluded who did not have morbidity assessed at the time of repeat operation and also if the amount of lung removed at reoperation was not specified (ie, chest wall resection with lung resection).
Identified were 203 patients who underwent lobectomy and later had a defined second lung resection that was assessed for the occurrence of any complications. Four patients had more than one repeat resection (and one had three repeat resections) for a total of 208 repeat lung resections. Seventeen patients underwent repeat resection within 30 days of the initial operation or during the same hospitalization for a diagnosis other than malignancy, such as hemoptysis, chest swelling, and pneumothorax. These were considered repeat operations done for complications of the initial operation and were excluded from the main body of analysis.
Patients that underwent operation within 30 days of initial operation and were coded as having a diagnosis of malignancy for the second operation were considered to have synchronous cancers and were retained in the main body of analysis. Thus, the remaining cohort of 186 patients was used in the bivariate and multivariate analyses of risk of reoperation in 191 repeat resections. In 18 patients, the postoperative outcome from their initial lobectomy was not recorded. These patients could not be used in the analysis of the effect of complications of initial operation as predictors of complication from reoperation, but they were included in all other analyses.
Statistical Analysis
For the purposes of the analyses, all variations of the repeat operations, such as sleeve resections, extrapleural resections, and lesser resections, were incorporated into one of the four major categories of lung resection, which were pneumonectomy, lobectomy, segmentectomy, or wedge resection. When a lesser resection was combined with a larger resection, such as wedge and lobectomy, the procedure was placed into the larger resection group. The different types of pneumonectomy, such as extrapleural, carinal, pneumonectomy, and completion pneumonectomy, were grouped into the single category of pneumonectomy for the purposes of this study.
A major complication of the initial lobectomy was defined by any of the following morbidities: pneumonia, reintubation, ventilator for more than 24 hours, pulmonary embolism, acute respiratory distress syndrome (ARDS), cardiac arrest with cardiopulmonary resuscitation, myocardial infarction, congestive heart failure, sepsis, deep wound infection, progressive renal insufficiency, or cerebrovascular accident.
In general, demographic, preoperative, and intraoperative factors were analyzed for their associations with morbidity and mortality if at least 5% of the total patients were represented in each category. The Fisher exact test was used in the bivariate analysis of mortality and morbidity with respect to the factors coded as categoric data. The Wilcoxon rank sum test was used to explore any differences in the distribution of continuous factors such as age and operative time between patients with an adverse outcome compared with those without. It was used to aid in choosing rational cutoffs for recoding each continuous variable into meaningful categories for analysis.
Factors that were significant at the p = 0.100 level in bivariate analysis of either mortality or morbidity were considered as independent variables in the multivariate analysis of each adverse outcome. Independently significant factors were identified by stepwise logistic regression with entry and exit criteria at the p = 0.100 level. Goodness-of-fit of the final logistic models was assessed by the Hosmer-Lemeshow statistic. A 95% confidence interval for the odds ratio was derived from the Wald confidence interval of each regression parameter. All p values were based on two-sided hypothesis tests. Computations were performed using SAS 9.1 (SAS Institute Inc, Cary, NC).
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
Repeat Operations
The 191 repeat resections consisted of 32 (17%) pneumonectomies, 84 (44%) lobectomies, 11 (6%) segmentectomies, and 64 (34%) wedge resections. During the operation, 15% of patients required transfusion. Median length of operation was 2.67 hours (range, 0.33 to 9.45 hours, with p = 0.011 compared with initial lobectomy. The median hospital stay was 7 days (range, 0 to 163 days), with p = 0.006 compared with initial lobectomy.
The 30-day mortality of patients undergoing repeat operation was 11%. The 30-day complication rate was 19%. Serious, frequent (>2% incidence) complications were pneumonia (9%), ventilator dependence exceeding 48 hours (6%), reintubation (8%), cardiac arrest (3%), cardiac dysrhythmia (3%), and sepsis (3%). A complete list of complications with an incidence exceeding 1% is listed in Table 2.
|
|
Factors associated with increased morbidity of repeat lung resection were wound classification, 67% in contaminated/infected cases versus 18% in clean/clean-contaminated cases (p = 0.014); blood transfusion, 15% if no transfusion versus 43% if one or more units transfused (p = 0.002); and total operative time, 31% if more than 4 hours versus 20% if 2 to 4 hours versus 11% if 2 hours or less (p = 0.038). Mortality and morbidity were not affected by a minimally invasive thoracoscopic approach versus standard thoracotomy or whether the postoperative diagnosis was cancer (International Classification of Disease, 9th revision, Clinical Modification, 140–239). Intraoperative factors and their correlation with morbidity and mortality are listed in Table 4.
|
Multivariable Analysis
Multivariable models for mortality and morbidity examined the joint effects of functional status, COPD, dyspnea, WBC, hematocrit, BUN, serum sodium, pneumonectomy, pneumonectomy/lobectomy versus lesser resections, wound status, transfusion, operation time, and complications of initial lobectomy. Only the intraoperative variables retained statistical significance in the multivariate model of mortality. Size of the odds ratios suggests these effects overwhelmed all the other factors in increasing the risk of death. Intraoperative variables associated with increased risk of death were contaminated/infected case, pneumonectomy, and intraoperative blood transfusion. (Table 5).
|
|
|
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Although these studies provide perioperative mortality and survival data, only one reported complications, but without risk analysis. In our report, we used a prospective database to analyze the high risks of repeat lung resection after lobectomy. Overall 30-day mortality was 11%, and 30-day morbidity was 19%. The mortality from repeat resection in this study is about double that reported in the mentioned single-institution retrospective studies. The percentage of patients undergoing reoperation pneumonectomy in our series (17%) is not higher than most other reports and thus does not account for the higher mortality rate. The overwhelming predominance of men in the VA system may account for some of the increased mortality rate.
The NSQIP database includes both high-volume tertiary centers and lower-volume centers, which may also explain the higher mortality rate than that described reports from single tertiary centers. Surgeon training, experience, and board certification may also account for some of the differences, although these variables are not recorded in the NSQIP database. We also believe that the prospective method of gathering perioperative data generally leads to more inclusive and more accurate follow-up data.
The perioperative mortality of patients undergoing lung resection after a previous lobectomy is two to three times that of initial lobectomy or pneumonectomy. A report by Harpole and colleagues [10] from the same VA NSQIP database published in 1999 on initial resection described a 4% mortality and 24% morbidity for lobectomy and a 11% mortality and 26% morbidity from pneumonectomy. This is in comparison with the 7% mortality and 21% morbidity for repeat lobectomy and 34% mortality and 28% morbidity for pneumonectomy after lobectomy in our study.
The initial lobectomy results of our repeat resection population had a much lower morbidity (13%) than the 24% reported by Harpole and colleagues [10]. The patients in this study were inherently healthier, as defined by low incidence of initial perioperative morbidity, than the typical lobectomy patient. This strongly suggests that the patients in the current study have been self-selected for reoperation. Clearly, the sickest patients were those that died after the initial lobectomy and are thus excluded from the repeat resection population. Moreover, patients that sustained severe, permanent, or even serious complications may have refused offers of repeat operation or may have not been offered repeat operation by their surgeon. The average age at repeat operation was only 13 months greater than at initial operation.
Whether the increased risk of repeat operation is due to the decrease in lung capacity, decreased volume of the pulmonary vascular bed, or increased difficulty of reoperation because of scarring is not clear. The independent increase in mortality seen with the need for intraoperative transfusion and the increase in morbidity seen with operative time suggests that increased difficulty of operation is at least partly the cause of that increased morbidity.
It is clear that the amount of lung resected during repeat operation has an effect on perioperative mortality, although not necessarily morbidity. The risk of death from pneumonectomy was nearly sixfold higher (34% versus 6%) than that for lesser resection (p < 0.01). Segmentectomy or wedge resection had one third the mortality (5% versus 15%) of lobectomy or pneumonectomy, with a trend toward statistical significance (p = 0.057). The mortality of pneumonectomy after lobectomy has traditionally been among the highest of all lung resections, with two recent reports describing a mortality rate of 20% [14, 15]. Our data strongly suggest that as long as negative margins can be achieved, lesser resections should be performed in preference to lobectomy or completion pneumonectomy for recurrent cancer.
Although the NSQIP data do not include survival data, it is difficult to imagine that any potential improvement in survival from completion pneumonectomy could compensate for its added perioperative mortality. Although the Lung Cancer Study Group Study [1] showed a decrease in recurrence rate and a borderline significant increase in survival with lobectomy versus lesser resection for first time resections, the threefold difference in perioperative mortality between lobectomy/pneumonectomy and limited resections seen in these repeat resections makes it unlikely that overall survival would be improved by repeat lobectomy compared with lesser resections.
The presence of active infection, as evidenced by a preoperative WBC exceeding 10,000/mm3 or a contaminated or infected case, was an independent predictor of postoperative morbidity and death. This has been noted in several reports, mostly in the risk of completion pneumonectomy. Miller and colleagues [14], in their report of 115 completion pneumonectomies, found a 26.3% risk of death for completion pneumonectomy done for benign disease, whereas that for malignant disease was 17.6%. In their meta-review of the literature, the overall perioperative mortality rate for benign disease was 18.1% versus 9.3% for malignant disease [14]. In another report of completion pneumonectomy, preoperative infection in the pleural space, most likely caused by bronchopleural fistula, was associated with a hospital mortality rate of 23% versus 11% for noninfectious resections [16]. Certainly, an infection in the lung parenchyma would also increase the morbidity and mortality.
Little association was found between complications as a result of initial lobectomy and subsequent morbidity during reoperation. An association of borderline significance (p = 0.084) was found between the presence of one or more complication during the initial lobectomy and the risk of death after repeat resection, with a mortality of 25% if one or more complications during prior lobectomy versus 10% mortality if no prior complication. This did not persist in multivariate analysis.
The interval between initial lobectomy and repeat lung resection did not affect perioperative morbidity or mortality. Excluding repeat resections performed for complications of initial resection, cutoffs of 1 month, 6 weeks, 3 months, 6 months, or 1 year between operations were not predictive of morbidity or mortality. Because of pain issues and the transient postoperative decline in lung function, many surgeons choose to wait a period of time before operating on a synchronous tumor after thoracotomy and lobectomy. These data suggest that this delay may not be strictly necessary. Likewise, no difference in risk was found between nonsmokers and prior smokers, and no correlation was found between the amount smoked (pack-years) and risk of reoperation.
The study has several limitations. First, it was conducted in the VA population, which is almost exclusively male. Several studies have shown male gender to be an independent predictor of morbidity during lung resection [17, 18]. This may partly explain the higher than previously described mortality from completion pneumonectomy.
Second, the NSQIP database is a general database designed to study patients undergoing a variety of surgical procedures, and does not, to date, collect specialty-specific information such as pulmonary function, side of operation, or cancer staging. Pulmonary function, especially the percent predicted postoperative forced expiratory volume in 1 second [19] and the diffusing capacity of the lung for carbon monoxide [17], has been shown to be among the best predictors of postoperative morbidity and mortality after initial lung resection. Their absence in the preoperative factors clearly limits the utility of this database in evaluating lung resection morbidity. This issue is currently being addressed by the development of specialty specific databases within the NSQIP database. Side of operation is important when attempting to evaluate the effect of scarring on risk of operation. In addition, repeat resections on the contralateral lung after ipsilateral lobectomy are more likely to be complicated by difficulty with single-lung ventilation. The role of these factors cannot be addressed with the current database.
Third, the accuracy of the database depends upon accurate input by surgeons and trained database nurses. Although there are separate codes for pneumonectomy and completion pneumonectomy, several observed inconsistencies and the absence of coding for side of procedure required the placement of these two separately coded procedures into a single pneumonectomy group for the purposes of analyses. There were 16 cases coded as completion pneumonectomy and 16 cases coded as pneumonectomy. The mortality and morbidity of the two procedures were similar (p = 1.000 for both comparisons).
Fourth, while the predictive validity of each model was good [The c-index for the morbidity model was 0.752. The c-index for the mortality model was 0.823. The final logistic models of mortality and morbidity provide a good level of fit based on the Hosmer-Lemeshow statistics of 0.046 with 2 degrees of freedom (p = 0.977) and 0.322 with 4 degrees of freedom (p = 0.988), respectively], there are relatively few outcome events per covariate analyzed. In our analysis, there were 10 covariates used in the analysis of 21 deaths and 7 covariates used in the analysis of 37 complications. With such a small sample size and number of events, there is instability in the selection of variables for the models.
In summary, the risk of repeat lung resection after lobectomy is significant, and in this database without preoperative lung function data, can be predicted by the presence of infection, need for intraoperative blood transfusion, total operative time, presence of dyspnea at rest or on minimal exertion, and performance of pneumonectomy versus a lesser resection.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
It is not surprising that the functional status or albumin as a sort of a marker of functional status was not significant, because I think you preselected these patients, because, frankly, they came back, they had had a previous lobe, and someone decided that they were fit enough to have another operation. So I think that is probably why the albumin had dropped out, you know, because these were probably relatively fit patients.
These are excellent data. They certainly show that you can do these. And to tell you the truth, I don’t think an 11% mortality is that extreme, considering these patients. You had a lot of completion pneumonectomies or pneumonectomies in there. But how are you going to use these data? Are you going to try to develop some sort of algorithm for physicians? Exactly what do you plan on doing with them? I think this is a nice study.
DR LINDEN: Well, one of the questions is, someone who has had a lobectomy and presents with a small peripheral cancer, without evidence of metastatic disease, what is the best operation for that patient? Is it a limited resection, such as a wedge resection or segmentectomy as opposed to a completion pneumonectomy? Again, we don’t have survival data here, but the difference in perioperative mortality is so great that I think it favors limited resections, whenever possible.
DR MARK I. BLOCK (Hollywood, FL): I agree with Dave. This is a very nice study and a great use of a terrific database. I have a couple of questions. First, I noticed that the mean interval between the lobectomy and the subsequent operation was 13 months, which seems awfully short. I would be interested in your interpretation.
DR LINDEN: That is a short interval, and that suggests that many of these cancers may be synchronous. We couldn’t separate which are which here, but that is what I would probably derive from that interval.
DR BLOCK: Part of that question is, were you able to look at the interval between the operations as a predictor? I am sorry, if you said it, I missed it.
DR LINDEN: We did. We did it as a continuous number and as separate points, that is, 6 weeks, 3 months, 6 months, and 1 year. At none of these points was the interval an independent predictor of either complication or mortality from repeat operation.
DR BLOCK: My last question is, were you able in this group database to look at whether these were contralateral versus ipsilateral? In other words, were these all completion pneumonectomies or did somebody actually do a pneumonectomy on the other side after they had done a lobectomy?
DR LINDEN: That is one of the weaknesses of this database. Side of procedure is not recorded. Let me say that we believe that the vast majority were completion pneumonectomy. In fact, looking at the differences in mortality between those coded as pneumonectomy and those coded as completion pneumonectomy, there was no statistical difference in the mortality in those subgroups, so we felt justified grouping these together.
DR BLOCK: Thank you very much. I enjoyed it.
DR JOHN E. MOORE (Atlanta, GA): Congratulations on an excellent paper. I wanted to ask you to elaborate a little bit about the deaths that were associated with the intraoperative transfusion. Others have noted this. Were these pulmonary deaths from ARDS or was it just something that happened to be noticed in looking through your data?
DR LINDEN: I don’t have information on individual deaths and causes of individual deaths. That wasn’t always listed. In our analysis, intraoperative factors overwhelmed all other factors in regards to predictors of mortality, and this suggests that difficulty of operation, and not necessarily the medical condition of the patient is responsible for perioperative mortality.
DR MOORE: So you are speculating that receiving the intraoperative transfusion was not the cause. It just occurred because of things that happened in the operating room.
DR LINDEN: Yes. We can’t tell from here, but that is what I might think.
DR ARA VAPORCIYAN (Houston, TX): I am intrigued by the white count greater than 10,000 as a risk factor or predictor. Is that because a significant proportion of patients were coming to the operating room with septic complications from a prior lobectomy, like a late empyema, or is that something you can’t distinguish from this database?
DR LINDEN: We excluded patients brought back to the operating room within 30 days of lobectomy with nonmalignant diagnostic codes from the main body of analysis, as these patients most likely underwent repeat resection as a complication of the initial lobectomy. The database does not contain the information to determine if resections later than 30 days from the initial lobectomy were for empyema.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  B. Planquette, F. Le Pimpec-Barthes, L. Trinquart, G. Meyer, M. Riquet, and O. Sanchez Early respiratory acidosis is a new risk factor for pneumonia after lung resection Interact CardioVasc Thorac Surg, March 1, 2012; 14(3): 244 - 248. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Viviano, M. Renius, J.-C. Ruckert, A. Bloch, C. Meisel, A. Harbeck-Seu, W. Boemke, M. Hensel, K.-D. Wernecke, and C. Spies Selective Neurogenic Blockade and Perioperative Immune Reactivity in Patients Undergoing Lung Resection Journal of International Medical Research, February 1, 2012; 40(1): 141 - 156. [Abstract] [PDF]
|
|
|
|
 |
|
 R. S. Blank, C. Hucklenbruch, K. K. Gurka, D. C. Scalzo, X.-Q. Wang, D. R. Jones, S. R. Tanner, and J. M. Jaeger Intraoperative Factors and the Risk of Respiratory Complications After Pneumonectomy Ann. Thorac. Surg., October 1, 2011; 92(4): 1188 - 1194. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Bernard, C. Rivera, P. B. Pages, P. E. Falcoz, E. Vicaut, and M. Dahan Risk model of in-hospital mortality after pulmonary resection for cancer: A national database of the French Society of Thoracic and Cardiovascular Surgery (Epithor) J. Thorac. Cardiovasc. Surg., February 1, 2011; 141(2): 449 - 458. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Brunelli, A. Charloux, C. T. Bolliger, G. Rocco, J-P. Sculier, G. Varela, M. Licker, M. K. Ferguson, C. Faivre-Finn, R. M. Huber, et al. ERS/ESTS clinical guidelines on fitness for radical therapy in lung cancer patients (surgery and chemo-radiotherapy) Eur. Respir. J., July 1, 2009; 34(1): 17 - 41. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Sienel, S. Dango, A. Kirschbaum, B. Cucuruz, W. Horth, C. Stremmel, and B. Passlick Sublobar resections in stage IA non-small cell lung cancer: segmentectomies result in significantly better cancer-related survival than wedge resections Eur J Cardiothorac Surg, April 1, 2008; 33(4): 728 - 734. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
