Experience With the Preoperative Invasive Measurement of Haemodynamic, Respiratory and Renal Function in 100 Elderly Patients Scheduled for Major Abdominal Surgery

P. OLDER* AND R. SMITH
Department of Anaesthesia and Intensive Care, Western General Hospital, Melbourne, Victoria

Recent literature suggests that clinical examination and routine preoperative tests may not be sufficient to detect the extend of serious cardiovascular and other abnormalities in elderly patients presenting for major surgery and that invasive monitoring may assist in detecting this subset of patients.¹  Other work is very suggestive that perioperative heamodynamic monitoring improves survival in patients undergoing surgery with a previous history of myocardial infarction.²  The Faculty of Anaesthetists of the Royal Australiasian College of Surgeons has stated that there is insufficient emphasis paid to the prevention of postoperative surgical problems and that prevention may represent a more cost-effective method of utilizing intensive care beds than treating elderly patients with established organ system failure.³  Invasive monitoring may itself miss another subset of patients that could be detected by exercise testing as suggested recently by Goldman at a meeting of the Faculty of Anaesthetists of the Royal College of Surgeons in London (Goldman L., personal communication).

In this study intensive care unit facilities were used for invasive preoperative evaluation of haemodynamic, pulmonary and renal function prior to surgery.  It was felt that this approach used in combination with postoperative monitoring should reduce both morbidity and mortality.

While ICU facilities were used at the time of the study because no other place in the hospital was suitable for invasive monitoring, it would be preferable to admit these patients preoperatively to an area set aside for this purpose.

We did not compare the clinical findings with the results of invasive monitoring, as was done by Professor Del Guercio.¹  In the author’s view invasive monitoring is complementary to, and not in competition with, a full clinical examination of the patient.  The intention was to quantify the range of physiological variables found in one hundred elderly patients preoperatively.  This would allow for identification of patients at particular risk and also early recognition and treatment of postoperative changes.

PATIENTS AND METHODS

PATIENTS

Having previously retrospectively examined information on all major abdominal surgery patients performed at our hospital over the three previous years we were able to identify the following high risk groups based on mortality and serious morbidity.

These were elderly patients undergoing:

      1.        abdominal aortic aneurysm surgery,
2.        other intra-abdominal vascular surgery,
3.        colo-rectal surgery,
4.        other major abdominal surgery.

This prospective study of one hundred consecutive elderly patients who presented for elective major abdominal surgery were divided into those groups as follows:

      1.        abdominal aortic aneurysm surgery (34 patients),
2.        other intra-abdominal vascular surgery (11 patients),
3.        colo-rectal surgery (40 patients),
4.        other major abdominal surgery (15 patients).

METHODS

The patients were seen before ICU admission by an anaesthetist at which time a full history was taken, clinical examination was performed and existing tests reviewed.  The nature of the operation, the procedures and postoperative plans ere explained to the patient and consent obtained.  The patients were admitted to ICU in the 24-hour period prior to the planned operation date.

Upon admission the history and clinical findings were confirmed.  A twelve lead ECG was obtained.  A silastic urinary catheter was inserted.  An arterial line was established.  A pulmonary artery flotation catheter was inserted, usually into the left subclavian vein.  The position of this catheter was confirmed with a chest X-ray.

All measurements were made using Hewlett-Packard 78532B monitors with the patients sitting at approximately 45 degrees.  The reference point for zero pulmonary artery pressure was established.  The transducer was attached to the patient and the skin position marked.  This allowed repositioning of the transducer to the same zero reference point postoperatively.  This was important as the anaesthetists were then able to resite the transducer ion the operating theatre.

Haemodynamic studies were performed several times to allow the patient to stabilise.  No sedatives were used prior to these studies and the patient was continued on drugs that had been prescribed previously.  Creatinine clearance was calculated several times using a four-hour time period and accurately measured catheter samples of urine.  Urinary sodium, potassium, creatinine, urea and osmolarity were measured.  A full blood biochemical profile was obtained.  Much of this data was necessary in order to calculate such parameters as creatinine clearance, fractional excretion of sodium, oxygen transport and consumption data and in order to calculate the APACHE score⁴ of our patients on a day-to-day basis.

All haemodynamic variables, blood gases (both arterial and mixed venous), relevant biochemical data, weight, height and patient temperature were stored in a computer using programs prospectively designed by one of the authors (P.Older) to calculate all the standard haemodynamic variables.  In addition oxygen transport and consumption, alveolar/arterial gradients, pulmonary artery and systemic oxygen content differences and intrapulmonary shunting were calculated.  Shifts in the haemoglobin dissociation curve for arterial and mixed venous points were plotted, displayed graphically and stored.  For reasons of conformity the algorithms used were the same as those described by Shoemaker in 1985.⁵  The conversion of oxygen tension to saturation was performed using a modification of the equation as described by Kelman in 1966.⁶  All ECGs were reported by a consultant cardiologist.

RESULTS

This study identified severe physiological disturbances in a large number of elderly patients presenting for major abdominal surgery.  It also highlighted the large range of preoperative values.  Because group statistics using means and standard deviations may not demonstrate the extent of individual variations from normal, frequency histograms were used to present the findings.

Haemodynamics

Figure 1 is a frequency histogram for the cardiac index of the whole group.  Guyton⁷ offers an age-related figure for a normal cardiac index of between 2.4l/min/m² to 2.5l/min/m² for ages between 60 and 80 years.  Eleven per cent of our patients had a cardiac index of 2.2l/min/m² or less.  This figure was used by Forrester as one of the criteria for placing patients in the worst subset of haemodynamic performance following myocardial infaction.⁸

Blood Pressure

The range of mean blood pressures is presented in Figure 2.  Fifteen per cent of patients had previously untreated hypertension as defined by a mean blood pressure of over 120 mmHg.  This corresponded to 14% of patients with a diastolic pressure of over 100 mmHg.

Oxygen consumption index

The putative ‘normal’ oxygen consumption index⁹ is 140 ml/min/m² but some authors quote a lower figure of 125 ml/min/m².  Figure 3 shows that 70% of our patients had oxygen consumption below 130 ml/min/m² and 17% below 100 ml/min/m².  The importance of this is discussed below.  In our patients the oxygen consumption rose from a mean of 121 ml/min/m² preoperatively to 174 ml/min/m² postoperatively, an increase of 44%.  The extent of this rise is dependent on the magnitude and possibly the duration of the procedure.¹¹

Renal function

Figure 4 is the frequency histogram for creatinine clearance.  Of these patients 19% had a creatinine clearance of 50 ml/min or less.  Regression analysis of creatinine clearance and cardiac index did not show a correlation that was of clinical value (r=0.32).  Figure 5 shows this clearly.

Respiratory function

Respiratory function was studied by FEV₁ and FVC and by calculation of pulmonary venous admixture, as arterial blood gases are of limited value in estimating the efficacy of gas exchange.¹²  All shunts were checked, when elevated, with differing inspired oxygen concentrations.

While most patients had values of less than 10% there were a small number of serious elevations, 10% of patients having a measured shunt equal to or exceeding 15%.  The distribution for this variable is shown in Figure 6.

As might be expected there was no correlation at all between intrapulmonary shunt and measured values for FVC and FEV₁.  In keeping with the poor respiratory status of some of the patients there were five patients with significant pulmonary artery hypertension as defined by a mean pulmonary artery pressure at rest greater than 25 mmHg.  This is shown in Figure 7.

 DISCUSSION

The pulmonary artery catheter was chosen as a preoperative tool as it enables continuous assessment of pulmonary pressures and regular measurement of filling pressures, cardiac output, oxygen delivery and oxygen consumption over the entire perioperative period.  Metabolic carts are able to measure oxygen consumption but are unable to measure oxygen delivery systems or pressures.  Various scanning techniques are able to assess left ventricular function but are unable to measure oxygen consumption, nor are they practical for use on a continuous basis.

Many of the problems detected could easily be approved by simple measures such as adequately rehydrating patients, adjusting drug dosages, etc.  Some of these defects could have been treated on the ward, had they been recognised.  Many other problems were not so easily overcome and could not have been quantified without invasive monitoring.  While it is possible to suspect poor left ventricular function clinically it is not possible to measure it.  In particular the problems of identifying cardiac disease in vasculopaths by clinical means is well recognised.¹³

This study confirmed the high incidence of serious physiological disturbances in the elderly found by Del Guercio.¹  It also highlighted another problem.  Many of the patients preoperatively had parameters which were below normal even for their age.  If purely postoperative measurements are made one may be misled into believing that a postoperative oxygen consumption of 150 ml/min/m² is not a significant increase in oxygen demand.  It would not be if the preoperative value had been 140 ml/min/m².  If the preoperative value was only 100 ml/min/m² then it represents a 50% increase in oxygen demand,  even if the oxygen extraction ratio increase to 30% this would still require a considerable increase in cardiac output.

The relationship between poor cardiac function and subsequent renal performance is well extablished.¹⁴  As one might expect the poor overall cardiac status of these patients be accompanied by poor overall renal function as expressed by creatinine clearance.  Figure 5 shows that while this statement might be true as a generalisation it was not a clinically reliable correlation.  An elevated serum creatinine was associated with a poor creatinine clearance, but a ‘normal’ serum creatinine did not guarantee a ‘normal’ creatinine clearance.

It the result of clinical and invasive evaluation suggested that the current risk of operation outweighed the potential benefit then postponement or cancellation of the case was discussed with the patient, the surgical unit and the cardiologist.  No criteria for postponement or cancellation of cases was laid done before the study.

Seven patients had operations postponed as haemodynamic or other problems could not be rectified in the time available.  As these patients were scheduled for elective surgery postponement allowed time for optimisation of physiological parameters without compromising the patient.  All of the postponed operations were subsequently performed without incident.

Of these seven patients, three were on a combination of calcium channel blocking agents and beta-adrenergic blockage as antihypertensive therapy.  All three of these operations were postponed due to bradycardia and cardiac output depression.  While operations was postponed alternative antihypertensive therapy was instituted using haemodynamic monitoring to evaluate the effect on cardiac output and oxygen transport.  While combined beta-adrenergic blockade and calcium channel blocker therapy may prove beneficial for some patients at home, we believe that it is not appropriate therapy for the metabolic stress of major surgery.  If there is significant reduction in left ventricular function, it is our current policy to cease or reduce the dose of the beat-blocking agent when it is being used in combination with calcium channel blocking agents.  Two other patients were on combinations or verapamil and digoxin with serum levels of digoxin exceeding 4.5 nanomol/l (normal 0.6-2.5).

Two cases were postponed due to complete heart block.  Pacing was instituted preoperatively using temporary percutaneous flow directed pacing wires.

After consultation two patients had major procedures changed to lesser procedures which were performed without incidence.

Six patients had their planned operations cancelled.  Of the cancelled operations two patients had abdominal aortic aneurysms with cardiac indices of 1.9 l/min/m² or less and were aged in excess of 80 years.  Two were chronic vasculopaths with histories of multiple previous myocardial infarcts who also had cardiac indices of 2.3l/min/m² or less and impairment of creatinine clearance.  One of these patients opted for surgery despite the risks involved and dies from a myocardial infarct on day 2.  The remaining two operations were cancelled for other reasons including severe pulmonary artery hypertension and poor left ventricular function.  The details of these patients are shown in Table 1.  They were followed up after discharge from hospital and their current status is also shown in Table 1.

  Three other patients with chronic vascular disease were identified and documented as being at very high risk from a combination of low cardiac index, a virtual failure to respond to volume loading or vasodilator therapy, chronic renal failure and malnutrition.  However they were in need of essential repeat surgery for vascular lesions.  All three died.

There were eight deaths in the entire group.  Mortality was 6% in the aneurysm group, 5% in the colo-rectal and 7% in the miscellaneous abdominal group.  Our initial study had identified non-aneurysm intra-abdominal vascular surgery as carrying a very high mortality.  Of the eleven patients in that group, every patient had previous infarcts, in fact two patients had two previous infarcts and one patient three.  All of them had been subjected to previous surgery and were re-presenting for essentially limb salvage procedures.  Three of the eleven were cancelled and of the remaining eight, three patients dies as described above.  A study published this year showed a mortality of 16% for limb salvage surgery in patients over 70 years of age.¹⁵  The causes of death were classified after discussion with the surgeon concerned and are shown in Table 2.  

There were no cardiovascular deaths in patients who were identified preoperatively as low risk.  We feel that it is currently not possible to accurately forecast the risk of sepsis, supervening in elective surgical patients, and multisystem failure secondary to sepsis was the cause of death in four patients.  While the deaths in these four patients were classified as from ‘other causes’, it is important to know that tow of these patients were being supported on ventilators due to cardiopulmonary problems for many days before their death.  No patient died of renal failure as the sole cause of death.

This study showed that it is possible to identify a major subset of patients at high risk for major surgery by a combination of clinical examination and invasive cardiovascular and renal investigation.  The mean duration of stay in ICU was four days which includes admitting the patient a day early.  The mortality rates of the main groups, i.e. aneurysms and major colo-rectal surgery, were low at 5-6%.  Many operations went ahead with full knowledge of the risk where the surgeon felt the surgery was essential or the risk of not operating was greater than that of the operation.  While we were able to identify all the patients who dies from non-surgical causes as being at high risk, we identified three cases at high risk who in fact did very well.  This means that more work needs to be done to further refine our discriminators.  Form recent work by Szlachcic¹⁶ it would appear that exercise VO₂ may well detect patients who are unlikely to survive for twelve months because of their existing cardiovascular disease.  To perform major surgery on this subset of patients is highly likely to lead to death early in the postoperative period.  We feel that such methods of evaluation warrant more study in the surgical population.

ACKNOWLEDGEMENTS

We wish to thank Mr. J. Epstein for his invaluable help with the manuscript.

REFERENCES

1.        Del Guercio L, Cohn J.  Monitoring operative risk in the elderly.  JAMA 1980; 243:1350-1355.

2.        Rao T, Jacobs K, Adel E.  Reinfarction following anaesthesia in patients with myocardial  infarction.  Anesthesiology 1983; 59:499-505.

3.        Minimum Standards for Intensive Care Units.  Faculty of Anaesthetists, Royal Australiasian College of Surgeons.  Bulletin P10, 1985.

4.        Knaus W, Draper E, Wagner D, Zimmerman J.  An evaluation of outcome from intensive care in major medical centres.  Annals Int Med 1986; 104:410-418.

5.        Shoemaker W, Bland R, Appel P.  Therapy of critically ill postoperative patients based on outcome prediction and prospective clinical trials.  Surg Clin North Am 1985; 4:811-833.

6.        Kelman G.  Digital computer subroutine for the conversion of oxygen tension to saturation.  J Appl Physiol 1966; 21:1375-1376.

7.        Guyton A.  Textbook of Medical Physiology, 7^th ed.  W. B. Saunders Company, 1986; 273.

8.        Forrester J, Diamond G, Chatterjee K.  Medical Therapy of acute myocardial infarction by application of hemodynamic subsets.  N Engl J Med 1976; 295:1356.

9.        Bland R, Shoemaker W.  Common physiologic patterns in general surgical patients. Surg Clin North Am 1985; 4:793-809.

10.     Waxman K.  Physiologic responses to operation in high risk surgical patients.  Surg Gynecol Obstet 1981; 152:633-638.

11.     Waxman K.  Hemodynamic and metabolic changes during and following operation.  Crit Care Clin 198; 3:241-250.

12.     Dantzker DR.  Cardiopulmonary Critical Care.  Dantzker DR, ed. Grune & Stratton Inc., Orlando, Florida 1986; 38-45.

13.     Boucher B, Brewster D, Darling C, Okada R, Strauss W, Pohost G.  Determination of cardiac risk by dipyridamole-thallium imaging before peripheral vascular surgery. N Engl J Med 1985; 312:389-394.

14.     Sladen R.  Effect of anesthesia and surgery on renal function.  Crit Care Clin 1987; 3:373-393.

15.     Griffith CD, Callum KG.  Limb salvage surgery in a district general hospital: factors affecting outcome.  Ann R Coll Surg Engl 1988; 70:95-98.

16.     Szlachcic J, Massie B, Kramer B, Topic N, Tuibau J.  Correlates and prognosis implication of exercise capacity in chronic congestive heart failure.  Am J Cardiol 1985; 55:1037-1042.