Learning from regret.

BACKGROUND: Death after surgery is infrequent but can be devastating for the surgeon. Surgeons may experience intense emotional reactions after a patient's death, reflecting on their part in the death and the patient's loss of life. Excessive rumination or feelings of regret may have lasting negative consequences, but these reactions may also facilitate learning for future decision-making. This qualitative study analysed surgeons' reflections on what might have been done differently before a patient's death and explored non-technical (cognitive and interpersonal) aspects of care as potential targets for improvement. METHODS: In Australia's Queensland Audit of Surgical Mortality, surgeons reflect on factors surrounding the death of patients in their care and respond to the open-ended question: in retrospect, would you have done anything differently? Framework analysis was applied to surgeons' responses to identify themes relating to non-technical aspects of care. RESULTS: Responses from 1214 surgeons were analysed. Two main themes were identified. Dilemmas and difficult decisions confirmed the uncertainty, complexity and situational pressures that often precede a surgical death; regret and empathy for patients featured in some responses. In the second main theme, communication matters, surgeons cited better communication, with patients, families, colleagues and at handover, as a source of reflective change to improve decision-making and reduce regret. CONCLUSION: Surgical decision-making involves uncertainty, and regret may occur after a patient's death. Enhancing the quality of communication with patients and peers in comprehensive assessment of the surgical patient may mitigate postdecision regret among surgeons.

Processes of care in surgical patients who died with hospital-acquired infections in Australian hospitals.

BACKGROUND: Infection may complicate surgical patients' hospital admission. The effect of hospital-acquired infections (HAIs) on processes of care among surgical patients who died is unknown. AIM: To investigate the effect of HAIs on processes of care in surgical patients who died in hospital. METHODS: Surgeon-recorded infection data extracted from a national Australian surgical mortality audit (2012-2016) were grouped into HAIs and no infection. The audit included all-age surgical patients, who died in hospital. Not all patients had surgery. Excluded from analysis were patients with community-acquired infection and those with missing timing of infection. Multivariate logistic regression was used to determine the adjusted effects of HAIs on the processes of care in these patients. Costs associated with HAIs were estimated. FINDINGS: One-fifth of surgical patients who died did so with an HAI (2242 out of 11,681; 19.2%). HAI patients had increased processes of care compared to those who died without infection: postoperative complications [51.0% vs 30.3%; adjusted odds ratio (aOR): 2.20; 95% confidence interval (CI): 1.98-2.45; P < 0.001]; unplanned reoperations (22.6% vs 10.9%; aOR: 2.38; 95% CI: 2.09-2.71; P < 0.001) and unplanned intensive care unit admission (29.3% vs 14.8%; aOR: 2.18; 95% CI: 1.94-2.45; P < 0.001). HAI patients had longer hospital admissions and greater hospital costs than those without infection. CONCLUSION: HAIs were associated with increased processes of care and costs in surgical patients who died; these outcomes need to be investigated in surgical patients who survive.

Predictors of mortality in patients with femoral neck fracture.

PURPOSE: To identify patient and procedural factors associated with in-hospital mortality following a femoral neck fracture. METHODS: Records of 598 female and 259 male consecutive patients aged 29 to 108 (median, 82) years admitted between 2010 and 2014 with femoral neck fracture were retrospectively reviewed to determine patient and procedural factors associated with inhospital mortality. RESULTS: 73% of patients were operated on within 48 hours of admission. The in-hospital mortality was 7.5%, with 2.1% occurring preoperatively and 5.4% postoperatively. Factors associated with increased preoperative mortality included being nonambulant prior to admission (p=0.015), residence in interim care (p=0.001) or low-level care (p=0.049), having synchronous fractures (p=0.001), and having a concurrent acute medical condition (p<0.001). Patient factors associated with increased in-hospital mortality included male gender (p=0.041), age >80 years (p=0.001), non-ambulatory status (p=0.015), residence in high-level care (p=0.031) or low-level care (p=0.018), American Society of Anesthesiologists grade 4 or 5 (p<0.001), weekend admission (p<0.001), and having an acute medical condition on admission (p<0.001). Procedural factors associated with increased in-hospital mortality included >96-hour delay to surgery from admission (p<0.001), surgery over the weekend (p=0.005), and surgical interventions other than total hip arthroplasty (p<0.05). CONCLUSION: Identification of patient and procedural factors can guide changes in best practice to decrease mortality following a femoral neck fracture.

Analysis of deficiencies in care following cholecystectomy.

INTRODUCTION: Failure To Rescue was first defined in patients who died due to a complication following (open) cholecystectomy but research into the relevant factors has been scarce. This study was designed to determine a chronological sequence of deficiencies in care. METHODS: Adult patients who died under the care of a surgeon following cholecystectomy in Queensland were identified from the Australian and New Zealand Audit of Surgical Mortality (ANZASM) database. RESULTS: Not unexpectedly, this is a high-risk patient population: median age of the 48 patients was 74.5 years and the median number of comorbidities and American Society of Anesthesiologists class was 4. Death occurred on postoperative day 6. Most deaths occurred at the end of the week. Over 80% of deaths followed emergency cholecystectomy. In almost half the patients, there were no deficiencies in care. Most common deficiency was during postoperative management (i.e. Failure To Rescue), however, significant deficiencies also arose prior to surgical admission; choice and timing of intervention as well as intraoperative decision-making. CONCLUSION: Surgeons who perform cholecystectomy need to be aware of the levels at which deficiencies arise given that many may be preventable.

Analysis of the causes and effects of delay before diagnosis using surgical mortality data.

BACKGROUND: The aim of the study was to assess the causes and effects of delay in diagnosis in surgical patients who died in 20 public hospitals participating in the Queensland Audit of Surgical Mortality (QASM) in Australia. METHODS: This was a retrospective cross-sectional analysis (June 2007 to December 2011) of deaths reported to QASM. Deaths were assigned to one of two groups (no delay or delay in diagnosis). Logistic regression was used to compare the association of delay with surgical complications, both overall and by surgical specialty. RESULTS: A total of 3139 deaths were reported. Diagnostic delay was reported in 293 (9·3 per cent). The primary cause of delay was attributed to diagnostic support services (41·7 per cent). Some 174 (13·8 per cent) of 1259 general (gastrointestinal) surgery patients experienced delayed diagnosis. Delay across all surgical specialties was associated with an increased risk of unplanned return to theatre (odds ratio (OR) 1·77, 95 per cent confidence interval 1·24 to 2·52), of being treated in intensive care (OR 1·71, 1·15 to 2·54) and of postoperative complications (OR 1·39, 1·05 to 1·85). CONCLUSION: General (gastrointestinal) surgery patients who experienced delayed diagnosis were at increased risk of postoperative complications.

Tottering mouse motor dysfunction is abolished on the Purkinje cell degeneration (pcd) mutant background.

Tottering (tg) mice inherit a recessive mutation of the calcium channel alpha 1A subunit gene, which encodes the pore-forming protein of P/Q-type voltage-sensitive calcium channels and is predominantly expressed in cerebellar granule and Purkinje neurons. The phenotypic consequences of the tottering mutation include ataxia, polyspike discharges, and an intermittent motor dysfunction best described as paroxysmal dystonia. These dystonic episodes induce c-fos mRNA expression in the cerebellar circuitry, including cerebellar granule and Purkinje neurons, deep cerebellar nuclei, and the postsynaptic targets of the deep nuclei. Cellular abnormalities associated with the mutation include hyperarborization of brainstem nucleus locus ceruleus axons and abnormal expression of L-type calcium channels in cerebellar Purkinje cells. Here, the role of these two distinct neural pathways in the expression of tottering mouse intermittent dystonia was assessed. Lesion of locus ceruleus axons with the neurotoxin N-(2-chloroethyl)-N-ethyl-2-bromobenzyl-amine (DSP-4) did not affect the frequency of tottering mouse dystonic episodes. In contrast, removal of cerebellar Purkinje cells with the Purkinje cell degeneration (pcd) mutation by generation of tg/tg; pcd/pcd double mutant mice completely eliminated tottering mouse dystonia. Further, the c-fos expression pattern of tg/tg; pcd/pcd double mutants following restraint was indistinguishable from that of wild-type mice, suggesting that the pcd lesion eliminated an essential link in this abnormal neural network. These data suggest that the cerebellar cortex, where the mutant gene is abundantly expressed, contributes to the expression of tottering mouse dystonic episodes.

False aneurysm following intracranial surgery.

A false aneurysm of the pericallosal artery formed following resection of a colloid cyst. The aneurysm presumably resulted from arterial damage during surgery and its enlargement may have been facilitated by the surgical defect in the corpus callosum. A spherical hyperdense lesion on postoperative CT may indicate a false aneurysm.

Primary (granulomatous) angiitis of the central nervous system with multiple aneurysms of spinal arteries. Case report.

A 55-year-old woman with primary (granulomatous) angiitis of the central nervous system in association with non-Hodgkin's lymphoma (mucosa-associated lymphoid tissue type) presented with an acute spinal subdural hemorrhage secondary to rupture of one of several fusiform inflammatory aneurysms of the spinal cord radicular arteries. The literature on hemorrhagic complications, aneurysms, and spinal cord involvement in granulomatous angiitis is reviewed. Recognition of granulomatous angiitis is important, as the condition may be responsive to immunosuppressive therapy.

Tumours of the brain and presence of antibodies to Toxoplasma gondii.

The possible association between prior infection with the protozoan Toxoplasma gondii and development of brain tumours was investigated as part of two Australian population-based case-control studies of adult brain tumours. One study, based in Adelaide, South Australia, collected blood from 73 subjects with glioma, 53 subjects with meningioma and 348 controls. The other study, based in Melbourne, Victoria, collected blood from 44 subjects with glioma and 67 controls. All tumours had been verified histologically. IgG antibodies to T. gondii were measured using Enzyme Linked Immunosorbent Assay (ELISA) techniques. In both the centre-specific and combined analyses, there was no difference between subjects with glioma and controls in the prevalence of antibody test-positivity (35% test-positive in glioma versus 33% in controls, age-, sex- and centre-adjusted odds ratio (OR) = 1.00, 95% confidence interval (CI): 0.64-1.56). In the Adelaide study, there was a statistically significant increased risk of meningioma associated with antibody test-positivity (47% test-positive in meningioma versus 31% in controls, P = 0.02, adjusted OR = 2.09, 95% CI: 1.14-3.83). Our results do not support the hypothesis that antibody positivity to T. gondii is a risk factor for glioma, but suggest that it might be associated with meningioma.

Extradural haematoma: trends in outcome over 35 years.

We have reviewed 35 years experience of extradural haemorrhage (EDH) in a large neurosurgical unit, based in two university hospitals, one dealing exclusively with children and the other a general hospital. A steady reduction in the mortality rate from 29 to 8.5% occurred during that period. A trend towards earlier diagnosis is noted and an increasing proportion of rural patients has been evident throughout the study period. During the time-period studied there were many significant developments: the establishment of a modern neurosurgical unit, the evolution of an intensive care unit, the availability of CT head scanning and the formal organization of rapid retrieval to service country areas. However, no single feature could be identified as the major contributor to falling mortality results. Clinical awareness and early diagnosis are the keys to successful management of EDH.

The relative effects of hypoxic hypoxia and carbon monoxide on brain function in rabbits.

New Zealand white rabbits were exposed to control conditions (n = 11), or to either a progressive hypoxic hypoxia produced by dilution of oxygen (O2) with nitrogen (n = 10) or a 1% carbon monoxide (CO) admixture for 15 min (n = 11). Both exposures caused a significant increase in cerebral blood flow (CBF) of up to 300% such that O2 delivery to the brain was unchanged. In the hypoxia group, a cortical somatosensory evoked response (CSER) was unaffected until the arterial O2 tension was below 20 mmHg. At this time, the rabbits became hypotensive, O2 delivery to the brain decreased dramatically and the CSER could not be elicited. In contrast, despite the maintenance of O2 delivery to the brain during and after the CO exposure, the CSER voltages were halved during the exposure and only recovered to about 80% of baseline subsequently. We conclude that the primary toxicity of CO to the brain in rabbits is not due to a reduction in O2 delivery.

Extradural malignant lymphoma associated with spinal cord compression.

Cord compression was noted in 26 patients with extradural malignant lymphoma. There were 19 cases of nonHodgkin's lymphoma and seven cases of Hodgkin's disease. Eighteen cases of non-Hodgkin's lymphoma had intermediate or high grade malignancy types according to the Working Formulation Lymphoma Classification. These patients represented 28% of all extradural malignant tumours seen at the Royal Adelaide Hospital (RAH) and 2.5% of all patients with malignant lymphoma, during an 11 year period. The patients were classified in three presenting groups: Group A, six patients with primary extranodal extradural lymphoma; Group B, nine patients with both extradural lymphoma and disseminated disease at initial presentation; and Group C, 11 patients who developed extradural lymphoma during the course of established disease. Only classification by groups appeared to affect survival time. Group A had the most favourable prognosis, with a 5 year survival of 83%. Five of the Group A patients had a relapse of lymphoma; four at distant sites and one in the retroperitoneum. Laminectomy was essential to provide a diagnosis in Group A patients. The most useful warning symptom of impending spinal cord compression was back and/or radicular pain, which preceded neurological deficit either by days or by up to 5 years. Plain spinal X-rays were abnormal in 64% of cases, emphasizing the value of this simple procedure.

Iatrogenic spinal subdural haematoma.

Spinal subdural haematoma is a rare condition, usually seen in association with lumbar puncture or a bleeding disorder. It carries a high morbidity and mortality, and successful treatment requires prompt surgical intervention. We present a case of mixed spinal subarachnoid and subdural haemorrhage complicating failed spinal anaesthesia combined with anti-coagulation in an elderly woman, together with a review of the literature.

Arachnoid cyst presenting as an extradural haematoma.

A case of extradural haematoma as the presenting feature of an arachnoid cyst is described. Haemorrhagic complications of arachnoid cysts are well known, but extradural haematoma is rare. A possible pathogenetic mechanism is proposed and the treatment of asymptomatic arachnoid cysts is discussed.

Primary central nervous system lymphoma. A report of 9 cases and review of the literature.

Primary central nervous system (CNS) lymphoma appears to be on the increase. This increase is real rather than due to changes in nosology, physician awareness or increase in high-risk groups. Diagnostic guidelines appear to be well defined and widely accepted. We report 9 cases of primary CNS lymphoma seen at our institution over a 5-year period. The clinical presentation was variable, but consisted mostly of headache, motor dysfunction, and memory and behavioural disturbance. Surgery is of value for diagnosis. In general, surgical decompression is not recommended. Glucocorticoids and radiotherapy appear effective in the short term. The role of chemotherapy is uncertain.

Extradural haemorrhage in infancy and childhood. A review of 35 years' experience in South Australia.

One hundred and two paediatric cases of extradural haemorrhage (EDH) were treated in Adelaide, South Australia, during the period 1954-1988; 10 were infants (0-2 years) and 92 were children (2-14 years). There were 9 deaths (mortality 8.8%). Long-term disabilities severe enough to interfere with school and/or employment were seen in 8 (7.8%) survivors. This relatively low number of adverse outcomes is partly an expression of a low incidence (5.9%) of associated intradural haematomas and few high-velocity impacts due to vehicular accidents. There is reason to believe that the results of treatment have improved in the decade 1977-1988. We attribute this in part to early diagnosis by computed tomography (CT), but a contributory factor may be earlier referrals from country centres to a paediatric trauma centre and rapid transfer, by air or road, by medical retrieval teams.

Synapses, axonal and dendritic patterns of GABA-immunoreactive neurons in human cerebral cortex.

Gamma-aminobutyric acid (GABA) containing neurons were characterized in human association cortex by a combination of Golgi impregnation and immunohistochemistry. Neurons were Golgi impregnated, gold toned, drawn and then classified on the basis of their dendritic and axonal arborization in layers I-VI. An antiserum to GABA was used to determine which of the impregnated neurons were immunopositive. Twenty-four GABA-positive cells were Golgi impregnated: 7 were bitufted with their dendrites predominantly radially oriented, and 17 were multipolar stellate cells. Three of the multipolar cells with large somata in the deep layers showed dendritic patterns similar to previously described basket cells. Nine of the multipolar stellate cells in layers III-VI showed characteristics of 'neurogliaform' neurons (Ramón y Cajal, 1899). The somata and the dendritic field of these cells were spherical, with diameters of about 10-15 microns and 200 microns, respectively. Their dendrites were smooth and slightly beaded. The axon collaterals were densely distributed in and around the dendritic field, in a spherical area with a diameter of at least 300 microns. The thin axon collaterals had only occasional 'en passant' swellings. Contacts between the axons of neurogliaform cells and the distal dendrites of Golgi-impregnated pyramidal cells were observed. Electron microscopic immunocytochemistry revealed that GABA immunopositive nerve terminals formed symmetric synaptic contacts with somata, with GABA immunonegative and immunopositive dendritic shafts and with dendritic spines. The results show that GABAergic neurons are heterogeneous with respect to their dendritic and axonal patterns. In addition to the chandelier and basket cells, which have been shown in animal studies to contain GABA, other cell types, most prominently the neurogliaform cells, terminating on the distal parts of neurons, also contain GABA and may have a inhibitory function. Many of the GABAergic terminals make synapses on dendritic spines and shafts in the human cerebral cortex.

Diffuse axonal injury in head trauma.

Diffuse axonal injury (DAI) as defined by detailed microscopic examination was found in 34 of 80 consecutive cases of head trauma surviving for a sufficient length of time to be clinically assessed by the Royal Adelaide Hospital Neurosurgery Unit. The findings indicate that there is a spectrum of axonal injury and that one third of cases of DAI recovered sufficiently to talk between the initial head injury producing coma and subsequent death. The macroscopic "marker" lesions in the corpus callosum and dorsolateral quadrants of the brainstem were present in only 15/34 of the cases and represented the most severe end of the spectrum of DAI.

Extradural haemorrhage: strategies for management in remote places.

A study of 109 cases of extradural haemorrhage (EDH) treated in South Australia over a period of 7 years showed that 35 cases (32.1 per cent) presented in country areas at considerable distances from a neurosurgical service: the mortality in these country cases was 22.9 per cent, comparing unfavourably with a mortality of 12.2 per cent in metropolitan cases. The country series contained a disproportionately large number of cases with multiple intracranial haemorrhages, which are known to have a poorer outcome; when these cases were excluded, the rural mortality (12.5 per cent) was only a little over the metropolitan mortality (9.7 per cent). These data suggest that it is possible to manage extradural haemorrhages successfully even in places remote from a neurosurgical centre, if communications and air transport are used effectively. However, it was found that emergency operations carried out in country hospitals were sometimes inadequate or done too late. Medical retrieval teams based on city hospitals were sent out on 15 occasions, either to assist a general surgeon to complete an emergency operation, or to provide intensive care during transfer to a neurosurgical unit. Osmotherapy (mannitol and/or frusemide) has been useful in gaining time for transfer; the choice between immediate operation and transfer may be difficult, and decisions should take transfer time, clinical state and rate of deterioration into account.