Background Transfusion guidelines refer to the taking into account of comorbidities. There are three main components to the decision-making process: the oxygen-carrying capacity (Hb); the ability of the patient to increase their cardiac output; and the patient’s ‘need’ for oxygen. The Hb concentration is the only factor that has a numerical value in the clinical setting.

Aim The aim of the study was to quantify comorbidities so that a model of transfusion strategies could be constructed.

Methods Senior staff (n = 29) in the Auckland Department of Anaesthesia completed a questionnaire on their beliefs about patients’ abilities to increase their cardiac output in the presence of various pathologies. In another questionnaire (n = 22) they were asked, obliquely, about the patients’ need of oxygen. They were asked what the likelihood of transfusion was, in the presence of certain pathologies, if the patient’s Hb was 95 g/l. The answers to these questions (a mark on a 100 mm visual analogue scale (VAS)) were used to rank the pathologies.

Results The raw data had a very broad dispersion but it was possible to rank the pathologies as shown in the table below.

Conclusions It is possible to rank comorbidities that are significant when deciding to transfuse a patient, but their value has yet to be determined.

Aim The aim of the study was to create a model of the decision to transfuse that takes into account patients’ comorbidities.

Methods The three main components of the transfusion decision-making process – the oxygen carrying capacity (Hb); the ability of the patient to increase their cardiac output (CO); and the patient’s ‘need’ for oxygen (O2n) – were combined in a series of 30 clinical scenarios. These scenarios were presented to the senior staff of the Auckland Department of Anaesthesia who were asked to make a decision about transfusion.

The Hb concentration is the only factor that has a numerical value in the clinical setting but using questionnaires the ranking of various pathologies with regard to CO and O2n had been achieved in a prior study.

The thirty scenarios were composed of three distinct groupings of clinical states. In 10 the Hb and CO values were, on average, constant, in another 10 the Hb and O2n values, and in the final 10 the CO and O2n values. Once these scenarios had been constructed the questions were then re-ordered using random numbers.

Results The response to these scenarios demonstrated correlation between the decision to transfuse and the Hb and the CO response but there was only a very weak correlation with O2n. From the data a graph was produced of Hb vs CO and the likelihood that the senior anaesthetists at Auckland Hospital would transfuse.

The black squares represent over 50% likelihood of transfusion, the circles less than 30%.

Conclusions A model of the decision can be made from Hb and CO values. Further investigation of the data may allow refinements to the borderline situations.

Aim Peri-operative myocardial ischaemia and infarction (PMI) are frequent complications of major vascular surgery. The criteria for the diagnosis of PMI have recently changed to incorporate troponin assays. In addition, some studies suggest that minor elevations in troponin T (TT) are associated with adverse long-term outcome. The aim of this study was to examine the impact of peri-operative elevations in TT levels on cardiac outcome of patients after aortic surgery.

Methods One hundred and thirty six patients underwent open abdominal aortic aneurysm (AAA) repair during the initial period of data collection. Troponin T was measured routinely on post-operative days 1, 3 and 5. The patients were divided into three groups:

During the follow-up period questionnaires were sent to the patients’ general practitioners, and information sourced from the hospital database to determine the incidence of adverse cardiac events and death.

Results Follow-up data were available on 85% of AAA patients. Long-term morbidity and mortality were highest in the group with positive TT assays at the time of surgery (Group 3 = 25%). Morbidity and mortality were also elevated in those with intermediate TT levels (Group 2 = 22%). Group 1 had the lowest complication rate (6%).

Conclusions Our study suggests an increased long-term risk associated with the diagnosis of PMI. In addition, we have demonstrated that intermediate elevations of TT appear to be associated with adverse cardiac outcome. The results suggest that TT may provide a useful tool for identifying patients at higher long-term risk of cardiovascular complications.

Aims As orthotopic liver transplantation (OLT) becomes a more widely performed procedure, questions have been raised regarding the ability of smaller units to maintain quality outcomes. We discuss the New Zealand Liver Transplant Unit (NZLTU) experience and our approach to this problem. We present the outcomes achieved for liver transplantation at Auckland Hospital in comparison with other liver transplantation centres.

Methods A review of the prospectively collected data from the NZLTU patient database from commencement of OLT in New Zealand in February 1998 to June 2002.

Results A total of 125 liver transplants were performed to June 2002. These include four paediatric transplants and 11 for fulminant hepatic failure. Over the four years our annual transplant rate has increased from 13 to 36. Thirty-day mortality is 0.8% and one-year survival is 92%, both of which figures are similar to those reported from large international centres and the UNOS and ELTR databases.

Indications for OLT were hepatitis B (32 patients), hepatitis C (20), primary sclerosing cholangitis (12), fulminant hepatic failure (11), primary biliary cirrhosis (9), alcoholic liver disease (10), combined heart-liver transplant (1) and other (30).

Eighty one patients are European, 16 Maori, 13 Polynesian, and 15 other. Mean transfusion volumes were RBC 2198 ml, FFP 3035 ml, platelet 567 ml and cryoprecipitate 97 ml.

Critical events included line problems 4%, peri-operative myocardial infarction 0.8%, pulmonary hypertension 2.4%, and arrhythmia 9%.There is no evidence of a learning curve.

Conclusions With a population of 3.6 million in New Zealand the anticipated requirement for liver transplantation equates to 36 adults and six children per annum, putting the NZLTU at the lower end of international unit size. Despite this, the NZLTU has acceptable rates of morbidity and mortality.

Aims Aortic aneurysm surgery is associated with significant morbidity and mortality. The aims of this study were to examine the outcome of aortic aneurysm surgery at Auckland Hospital and to identify any factors associated with adverse outcome. As cardiac complications are the most common cause of death we specifically examined the rate of peri-operative myocardial infarction (PMI) using troponin T assay.

Methods All patients undergoing abdominal aortic aneurysm (AAA) repair between 1 June 1999 and 30 September 2000 were prospectively enrolled. Data included demographic information, comorbidities, pre-operative assessment and investigations. The primary outcome measure was mortality. Additional outcome measures included PMI and other peri-operative medical and surgical complications. PMI was diagnosed when two of the following three criteria were present: a clinical event suggestive of an MI, a troponin T level >0.1, or a new ECG change consistent with an MI.

Results One hundred and thirty six patients undergoing AAA repair were identified. Mortality for the 89 non-ruptured AAA patients was 5.6%. Risk factors for death included a diagnosis of diabetes, intra-operative arrhythmia, increased post-operative troponin T, post-operative congestive heart failure (CHF), the need for re-operation and increased length of stay in intensive care.

Mortality for the 47 ruptured AAAs was 38%. Pre-operative risk factors for death included a lower pre-operative Hb level, increased blood transfusion requirements, diabetes mellitus, lower functional status (METS), and intubation prior to arrival in theatre. Intra-operative and post-operative factors associated with death included increased medical and surgical complications. PMI was diagnosed in 24% of cases. Risk factors included diabetes, use of dopamine, coagulopathy, increased number of operations, post-operative CHF, and increased length of stay.

Conclusions Outcome from AAA surgery in our institution is comparable with results from other published series. However, the rate of PMI is significantly higher than previously reported and reflects the increased sensitivity of the troponin assay in diagnosing myocardial damage.

Background Environmental standards for trace gas exposure in operating theatres measure ambient gas concentrations by sampling the room air. This can be done by either single-point measure or a timed-exposure measure. These methods have the disadvantage of being difficult to translate into the implications for an individual’s peak exposure.

Aim This study aimed to assess the feasibility of end tidal gas sampling of anaesthetists’ breath during gaseous induction of anaesthesia at Christchurch Hospital.

Methods The breath samples were obtained by asking anaesthetists to blow the second half of a tidal breath into a mylar gas collection bag. Samples were collected prior to and after a gaseous induction by the anaesthetist under test. The samples were analysed for sevoflurane concentration using a selective ion flow tube (SIFT) mass spectrometer and for CO2 using infrared spectrometry. This study is ongoing and data from the first 12 samples are presented

Results The range of pCO2 found was 31 to 42 mmHg. The peak sevoflurane concentration found post-induction was 58 parts per million (OSH limit = 2 ppm).

Discussion The technique appeared reliable in returning samples containing a high proportion of alveolar gas, as judged by CO2 content. The measurement of CO2 allows the sevoflurane concentration to be normalised to allow for variations in sampling technique. This method of exposure testing may have advantages when trying to assess single-point exposure of anaesthetists to trace gas concentrations.

Background Anaesthesia involves the processing of large amounts of information. Typically anaesthesia data are stable, with noise, but the underlying trend may change. One task of the anaesthetist is to detect changes promptly and reliably. One manner of presenting data in a way that may help detection of change is to include the history of the parameter. This can be done in a variety of ways including a graphical trend display. Although many ‘know’ that a trend display helps detection of changes this has not been conclusively demonstrated. The aim of this study was to examine the effect of the presence of a graphical trend on the speed of detection of changes in simulated data.

Methods Ten anaesthetists each viewed 30 simulations with 4 parameters displayed. In a random 50% of the simulations the current values only of the parameters were displayed, while in the remainder the history of the parameters were graphed against time. A computer model generated values for the parameters which, after a period of stability, changed to a new random value that was at least 10 units away from the initial value. The time constant for the change was a random number with a rectangular distribution between three and ten minutes. Parameters were updated at one-second intervals.

Subjects were asked to indicate when they thought a ‘significant change’ was occurring. A separate, clearly marked key was used for each parameter and direction of change. The time between the onset of the change and its detection was recorded. Pooled means and standard deviations were calculated and compared using a paired t-test.

Results Data from one subject were incomplete and were not included in the analysis. All subjects detected changes faster with a trend display (mean delay 13.1 cycles, SD 1.9 cycles) than without (mean 15.3, SD 2.0), p = 0.002. Changes were detected slightly more accurately without the trend display (47.7 (SD 3.1) vs 44.7 (SD 3.3)) but this difference was not statistically significant.

Conclusions In this model a graphical display of the history of a parameter led to significantly faster detection of changes but had no significant effect on the accuracy of detection. The evaluation system was a useful tool and further studies are planned.