| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Eur J Cardiothorac Surg 2006;30:126-131
© 2006 Elsevier Science NL
a Department of Cardiothoracic Surgery, The Cardiothoracic Centre-Liverpool, Thomas Drive, Liverpool L14 3PE, United Kingdom
b Department of Research and Development, The Cardiothoracic Centre-Liverpool, United Kingdom
Received 24 September 2005; received in revised form 6 March 2006; accepted 17 March 2006.
* Corresponding author. Tel.: +44 151 228 1616; fax: +44 151 220 8573. (Email: reddylcs{at}aol.com).
 |
Abstract |
|---|
 Top Abstract 1. Introduction2. Material and methods3. Results4. DiscussionAppendix AReferences |
|---|
Key Words: Off pump • CABG • Respiratory risk • Propensity matching • Ventilation • Alveolar arterial gradient
|
1. Introduction |
|---|
|
TopAbstract1. Introduction2. Material and methods3. Results4. DiscussionAppendix AReferences |
|---|
|
2. Material and methods |
|---|
|
TopAbstract1. Introduction2. Material and methods3. Results4. DiscussionAppendix AReferences |
|---|
Definitions and data collection methods have been previously published [8] and are available from the website http://www.nwheartaudit.nhs.uk/. Data was collected prospectively during the patient's admission as part of routine clinical practice and entered into our cardiac surgery registry.
2.2 Respiratory risk stratification
Respiratory risk stratification was performed on the propensity-matched groups (see Section 2.6) using the following variables: (1) Forced expiratory volume per second (FEV1) less than 65% of predicted, (2) patients aged more than 75 years, (3) history of current smoking, (4) history of NYHA class IV dyspnoea in combination with respiratory bronchodilator medication, and (5) a body mass index greater than 40 kg/m2. Patients with two or more of the above criteria were classified as high respiratory risk.
2.3 Criteria for comparison of off-pump effect
The length of post-operative ventilation was identified as the primary outcome to compare the difference between off-pump and on-pump patients. Prolonged ventilation was defined as the requirement for intermittent positive pressure ventilation for more than 48 h. This time period was chosen to differentiate those patients requiring ventilation either because of circulatory instability or neurological events that may preclude early extubation. Secondary outcomes collected included re-intubation, readmission to ICU, inotropic and intra-aortic balloon pump support, dialysis support, re-operation, and cerebrovascular accidents.
2.4 Alveolar–arterial gradient estimation
Alveolar–arterial gradients (A–a gradient) were calculated using the equation: A–a gradient = PAO2
– PaO2, where PAO2 is the partial pressure of oxygen in the alveoli and PaO2 is the partial pressure of oxygen in the arterial blood gas [9,10]. PAO2 was calculated from the equation: FiO2(760–47) – (PaCO2/0.8). FiO2 is the fractional inspired oxygen concentration, 760 is the atmospheric pressure, 47 is the partial pressure of water vapour at 37 °C, PaCO2 is the partial pressure of carbon dioxide in blood gas and 0.8 is the normal respiratory quotient at a normal dietary metabolic rate.
A–a gradients were calculated at the following time intervals (a) at the time of admission to the ITU, (b) 2 h, (c) 4 h, (d) 8 h, (e) 12 h, (f) 24 h and (g) at the time of extubation. Differences in A–a gradients were analysed as appropriate between both the groups.
2.5 Anaesthesia and ventilation management
Anaesthetic management of both groups of patients was similar and changes were made to suite off pump technique as appropriate. All the patients received fentanyl based induction and were on 1% isoflorane maintenance. Patients in the on-pump group remained unventilated during cardiopulmonary bypass. In off-pump group, patients were covered with warming blanket and were given warm intravenous fluids during the procedure. Postural heamodynamic changes in the off-pump group were managed by judicious use of intravenous fluids, vasoconstrictors and short acting beta-blockers and or vasodilators to facilitate the proximal graft anastamosis.
Both groups of patients received identical ventilation and post-operative care. The criteria for extubation were: that the patients were haemodynamically stable, drainage less than 50 ml/h, fully awake and able to move all limbs to command, PO2 more than 12 kPa on 0.5% FiO2, base deficit less than –3, tympanic temperature more than 36 °C, and respiratory rate more than 10 per minute. Primary decision for extubation was made by the nursing staff and any deviation from the norm is assessed by the anaesthetist and appropriate intervention made to address individual patients’ requirements.
Patients who required prolonged ventilation were assessed by the senior anaesthetist on a daily basis and a ventilation weaning protocol suggested. Adjustments if required were made by the medical staff during the day to suite the clinical situation.
2.6 Statistical analysis
We propensity-matched off-pump patients with unique on-pump patients to account for case mix differences [11]. To achieve this, logistic regression was used to develop a propensity score for off-pump group membership [12]. Table 1
lists the variables included in logistic regression to develop propensity scores. The C statistic, which is equivalent to receiver-operating characteristic curve (ROC curve), for this model was 0.74. Off-pump patients were matched with on-pump patients who had an identical five-digit propensity score using the greedy-match technique. If this could not be done, we proceeded to a 4-, 3-, 2-, or 1-digit match [13].
|
|
3. Results |
|---|
|
TopAbstract1. Introduction2. Material and methods3. Results4. DiscussionAppendix AReferences |
|---|
Respiratory risk stratification resulted in 73 off-pump and 55 on-pump high-risk patients (Table 2 ) available for analysis. In the off-pump group, four (5.5%) patients had more than two selection criteria, compared to one (1.8%) patient in the on-pump (p = 0.29) group. There were more patients with FEV1 < 65% in the off-pump group compared to the on-pump group [65 (89%) vs 40 (72.7%)], (p = 0.017). The pre-operative characteristics of these high respiratory risk patients are shown in Table 3 . The median logistic EuroSCORE for off-pump patients was 4.9 (25th and 75th percentiles: 3.3–8.4) compared to 4.2 (25th and 75th percentiles: 2.8–7.8), which was not statistical significant different (p = 0.25). There was no inter surgeon differences in terms of number of patients operated or the risk profiles of the patients, excluding the surgeon as a possible variable influencing the outcomes.
|
|
0.001. The median ventilation time was significantly shorter for off-pump patients (7 h [25th and 75th percentiles: 5–14] vs 12 h [25th and 75th percentiles: 7–18], p = 0.003). The distribution of ventilation times between off-pump and on-pump patients is shown in Fig. 1 . In the off-pump group, three (4.1%) patients required prolonged ventilation of more than 48 h, compared to eight (14.6%) patients in the on-pump group (p = 0.037). In-hospital mortality was similar in both groups (two patients (2.7%) off-pump vs three patients (5.4%) on-pump, p = 0.43). Table 4 shows other in-hospital outcomes between off-pump and on-pump patients.
|
|
|
|
|
4. Discussion |
|---|
|
TopAbstract1. Introduction2. Material and methods3. Results4. DiscussionAppendix AReferences |
|---|
Respiratory dysfunction following cardiac surgery is multifactorial. General anaesthesia, sedation, sternotomy, pleural opening, volume shifts between systemic and pulmonary circulations and altered compliance of right and left ventricle are common for both on-pump and off-pump patients. Patients subjected to cardiopulmonary bypass have an additional insult from hypothermia, unventilated lungs and the inflammatory response. Studies have shown increased free radical generation and evidence of alveolar basement membrane injury suggesting that this inflammatory response contributes to the ultra structural changes seen in the lung. In our study, the observed higher A–a gradient in the on-pump group was noticeable only at the time of admission to the ITU. This seems to suggest that the inflammatory component is either negligible or non-existent. However, the favourable gradient in the off-pump group, although not significant, suggests better physiological homeostasis. In the on-pump group, the slightly worse A–a gradients are more likely to be the result of the effects of hypothermia and atelectasis. Hypothermia is known to shift the oxygen haemoglobin saturation curve to the left thus impeding the oxygen transfer at the alveolar and arterial interface accounting for some of the increased gradient observed in the on-pump group. Atelectasis following hypothermic cardiopulmonary bypass was shown to be a major cause of hypoxemia and intrapulmonary shunting by Magnunsson et al. [14] with their experimental work on pigs. Atelectasis of sub segmental bronchioles is difficult to recognize clinically or radiologically. In our practice, a positive end expiratory pressure (PEEP) of 5 cm H2O is routinely used in all our post-CABG patients to facilitate lung expansion. Depending on the serial blood gases and the lung compliance, PEEP levels are manipulated to improve the gas exchange. Unfortunately we did not account for PEEP manipulation and have no data for intrapulmonary shunting. However, it seems likely that atelectasis may account for the short-lived difference in A–a gradients between on-pump and off-pump patients. It is our policy to maintain sedation in all patients (on-pump and off-pump) until core temperature is stable prior to extubation. Not surprisingly, off-pump patients required less sedation. This and the shorter overall operation times must contribute to earlier extubation.
Previous studies by Cox et al. [5] reported similar differences in the A–a gradients between off and on-pump patients in the early post-operative period. Their study also concluded that cardiopulmonary bypass had little or no effect on A–a gradients. However, their study was limited by small numbers and lack of respiratory risk stratification. Ventilation time as a clinical measure was investigated by Guler et al. [3] and reported that off-pump patients had shorter mean ventilation times. They also noted that on-pump patients had a worse FEV1 on the seventh post-operative day and a higher incidence of atelectasis. They also failed to notice any significant differences on PO2 and PCO2 levels at various post-operative periods. Their study was also limited by much small numbers with less than 20 patients in each group. Montes et al. [6] also investigated the effect of off-pump surgery on patients with known pulmonary dysfunction. They concluded that PaO2/FiO2 ratios were comparable between both the groups. Their study reports a small but noticeable ratio in favour of off-pump patients in the early post-operative period. Another study by Staton et al. [15] also showed shorter ventilation times and slightly improved A–a gradients in off-pump patients. The effect of off-pump surgery on pulmonary function tests and PO2 and PCO2 was studied by Cimen et al. [16]. However, they failed to observe any difference between both the groups throughout the post-operative period.
There are some limitations, which need to be considered when drawing conclusions from this report. The first is that it is an observational study and therefore could have some degree of selection bias. The retrospective nature of the study cannot account for unknown variables affecting the outcomes that are not correlated strongly with the variables used in the propensity matching. Although, propensity score matching is no substitute for a properly designed randomised control trial, retrospective comparisons with propensity matching are more versatile and may be more widely acceptable than randomised control trials [14]. This combined with robust respiratory risk stratification offers strength to our study. Another limitation is that the lack of pre-operative A–a gradient data hindered the assessment of the gradient shift between the pre- and post-operative period in both groups. A relatively small sample size is another limitation of this study. However, a highly significant difference in the outcome variable along with the fact that this difference was noted despite the off pump group being slightly of a higher risk make type I error unlikely.
In conclusion, this study shows that patients undergoing off-pump CABG require less post-operative ventilatory support. This was more evident in current smokers and patients with FEV1 less than 65% predicted. There is a small but noticeable A–a gradient in favour of off-pump patients in the early post-operative period. The observed A–a gradient and beneficial effects are more likely to be due to avoidance of hypothermia and post-operative atelectasis combined with shorter operating times and length of post-operative sedation.
|
Appendix A |
|---|
|
TopAbstract1. Introduction2. Material and methods3. Results4. DiscussionAppendix AReferences |
|---|
Dr G. Bolotin (Tel Aviv, Israel): Very important results. However, regarding the extubation time, in a lot of units, as in ours, this is influenced first by the decision of the surgeon, and sometimes off-pump patients tend to be extubated earlier just because of the procedure they underwent, a surgical decision, and from other variables, such as bleeding, that is probably less and may delay the on-pump patient. Can you elaborate on this?
Dr Reddy : Yes, it is also our experience that off-pump patients get extubated early. The reasons for that are two-fold. We have a fairly rigid protocol in our ITU for extubation, which are independent of the surgeon or the anaesthetist. This may also be true in most other units. Secondly, patients have to be normothermic, cardiovascularly stable and not bleeding for a patient to be extubated. Off-pump patients are comparatively normothermic and also bleed less. This instinctively makes them receive less sedation. These we feel are some of the factors that offer benefits to the off-pump group and to the high-risk groups. The A–a gradient data becomes important to explain these clinical outcomes.
Dr R. Lorusso (Brescia, Italy): Just to link to the previous discussion, I think that also the alveolar difference, the gradient, can be very much influenced by many factors, and your data didn’t show statistical significance, first of all. At least I was taught that whatever is not significant is just, I mean, as similar as chance. And the sedation can very much affect, again, the efficiency of ventilation. So I think you have to really do a blind study for the people who are going to extubate the patients and to have the same degree of intubation time and then you reassess the alveolar gradient. It could be really important to understand whether there is exactly a clear difference between the two techniques, otherwise I think it would be very difficult to draw conclusive information and data.
Dr Reddy : Yes, I certainly agree with all the criticism levelled against the methodology we adopted. This is indeed a retrospective study. But, perhaps, I could satisfy your criticism on a couple of these issues. The whole intention of calculating Alveolar–arterial gradient is to look for presumed and hypothesized inflammatory component induced by the cardiopulmonary bypass. There are studies in the literature that showed the presence of alveolar basement collagen breakdown product, 7S protein. If one accepts that theory, one would presume that it would increase the Alveolar–arterial gradient. This may indeed be true in one or two cases we see in a year who are unfortunate enough to develop a full-blown ARDS. It would not be possible to include these rare occurrences in a regular model. For a day-to-day practice, I think our study holds true. The slight difference in the A–a gradient between the groups at the time of admission to the ITU is better explained by the fact that on-pump patients tend to be hypothermic and as we all know, hypothermia shifts the oxygen-haemoglobin dissociation curve to the left and this would contribute to the A–a gradient in the on-pump group. There is also enough evidence in the literature to show that when a patient is put on cardiopulmonary bypass and the lungs are collapsed, there is always an element of subsegmental collapse and intrapulomonary shunting. This would add further to the A–a gradient. I think these are the two main factors that contribute to the A–a differences noticed rather than the inflammatory component. This in my view offers the best explanation.
I certainly understand the validity of the argument to do a prospective, double blind randomized study with pre-operative A–a gradient calculation. This remains one of the limitations of our study.
However, regarding the pre-operative A–a gradients, let us consider two hypothetical possibilities. Through out my presentation, I have shown that in this study, patients in the off-pump group were of slightly higher risk than the on-pump group. If these patients were to have worse pre-op A–a gradients than in the post-op, that would further substantiate the conclusions we drew. If we consider that possibility that the pre-operative gradient in off-pump group was better than the post-op, it would still mean that the same difference between the groups would remain and the conclusions we drew would still be valid. This will not contradict our conclusions simply because we would not know how these patients would have behaved under cardiopulmonary bypass. Under ideal scientific circumstances, if one were to conduct this experiment in the lab, we would ideally need to have both the groups of patients undergo both off-pump and on-pump surgery and look for the A–a gradient differences. As we all know that would not be possible in the real world. I still feel that this data gives us answers applicable to every day practice.
Dr Lorusso : I would suggest that you take the patients with the same intubation time and you do the difference with the alveolar gradient, and I think we could just skip all the possible criticisms, because with the same intubation time, if you have a difference, that could be really good information.
Dr Reddy : I appreciate your argument. This study was designed with the extubation time as the primary outcome variable. To compare the patients who had the same extubation times would contradict and contravene the primary aim of this study in evaluating the benefit of off-pump surgery. For the sake of argument, one could look at the groups of patients who had similar extubation times. We have in fact collected the data up to 48 h and till the extubation time. Most of our patients got extubated between 6 and 8 h leaving much smaller numbers for later periods. This prevented us from doing any meaningful statistics. For that reason, I did not show this data. When we looked at these later extubations, they were very similar without any differences between the groups. I hope that answers your question.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Presented at the joint 19th Annual Meeting of the European Association for Cardio-thoracic Surgery and the 13th Annual Meeting of the European Society of Thoracic Surgeons, Barcelona, Spain, September 25–28, 2005. |
References |
|---|
|
TopAbstract1. Introduction2. Material and methods3. Results4. DiscussionAppendix AReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Rukosujew, S. Klotz, C. Reitz, W. Gogarten, H. Welp, and H. H. Scheld Patients and complication with off-pump vs. on-pump cardiac surgery a single surgeon experience Interactive CardioVascular and Thoracic Surgery, December 1, 2007; 6(6): 768 - 771. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. C. Austin Propensity-score matching in the cardiovascular surgery literature from 2004 to 2006: a systematic review and suggestions for improvement. J. Thorac. Cardiovasc. Surg., November 1, 2007; 134(5): 1128 - 1135. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. L.C. Reddy, A. D. Grayson, E. M. Griffiths, D. M. Pullan, and A. Rashid Logistic Risk Model for Prolonged Ventilation After Adult Cardiac Surgery Ann. Thorac. Surg., August 1, 2007; 84(2): 528 - 536. [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 |
