Peter Roffey

A Critical Analysis of Perioperative Mortality From Radical Cystectomy

PURPOSE Operative mortality from radical cystectomy has decreased as a result of improvements in surgical and anesthetic care. We reviewed the perioperative deaths from a large group of patients treated with radical cystectomy for primary bladder cancer. MATERIALS AND METHODS All perioperative mortalities from radical cystectomy were identified from a single high volume institution. The medical records were reviewed to assess the cause of death as well as possible contributing factors. RESULTS From August 1971 to December 2001, 1,359 patients with primary bladder cancer were treated with radical cystectomy and pelvic iliac lymphadenectomy at our institution. Of these patients, 27 (2%) died within 30 days of surgery or before discharge from hospital. Median patient age at surgery was 67 years (range 47 to 78) and males accounted for 81% of the patients. The median time to death was 28 days from cystectomy (range 0 to 80). Most deaths were cardiovascular related (including acute myocardial infarction, cerebrovascular accident, arterial thrombosis) or due to septic complications with resulting multi-organ system failure, followed by pulmonary embolism, hepatic failure and hemorrhage. Septic related mortality was most often associated with postoperative urine or bowel leak. While most deaths occurred before hospital discharge, 2 patients died at home due to a late pulmonary embolus. No association was seen between pathological stage or type of urinary diversion and mortality. CONCLUSIONS Perioperative mortality from radical cystectomy is low in this group of patients. Most deaths are due to cardiovascular or septic complications. Careful patient selection and meticulous surgical technique may help decrease the incidence of perioperative mortality.

Intraoperative evaluation of tissue perfusion in high-risk patients by invasive and noninvasive hemodynamic monitoring.

OBJECTIVE Although invasive monitoring has not been effective in late stages after organ failure has occurred, early postoperative monitoring revealed differences in survivor and nonsurvivor patterns and provided goals for improving outcome. We searched for the earliest divergence of survivor and nonsurvivor circulatory changes as an approach to earlier preventive therapy. The aim was to describe the intraoperative time course of circulatory dysfunction in survivors and nonsurvivors among high-risk elective surgery patients using both the thermodilution pulmonary artery catheter (PAC) and multicomponent noninvasive monitoring. DESIGN Prospective intraoperative description of circulatory dysfunction. SETTING University-run county hospital. PATIENTS Two hundred nine consecutively monitored high-risk elective surgery patients. MEASUREMENTS AND MAIN RESULTS We evaluated the data of high-risk elective surgery patients using both PAC and multicomponent noninvasive monitoring. The latter consisted of the following: a) an improved bioimpedance method for estimating cardiac output; b) the standard pulse oximetry to screen for pulmonary problems; c) transcutaneous oxygen and carbon dioxide tension sensors to evaluate tissue perfusion; and d) routine noninvasive blood pressure and heart rate. The current noninvasive impedance cardiac output estimations closely approximated those of the thermodilution method; r2 = .74, p < .001; the precision and bias was -0.124 +/- 0.75 L/min/m2. Outcome measures included intraoperative description of circulatory patterns of high-risk surgical patients who survived compared with nonsurvivors. Hypotension, low cardiac index, arterial hemoglobin desaturation, low transcutaneous oxygen, high transcutaneous carbon dioxide tensions, low oxygen delivery, and low oxygen consumption developed intraoperatively gradually over time; the abnormalities were more pronounced in the nonsurvivors than in the survivors. CONCLUSIONS The survivors had slightly higher mean arterial pressure, cardiac index, and mixed venous oxygen saturation, as well as significantly higher oxygen delivery, oxygen consumption, transcutaneous oxygen tension, and transcutaneous oxygen tension/FIO2 ratios, than did the nonsurvivors. The data suggest that blood flow, oxygen delivery, and tissue oxygenation of the nonsurvivors became inadequate toward the end of the operation. Noninvasive monitoring provides similar information to that of the PAC; both approaches revealed low-flow and poor tissue perfusion that were worse in the nonsurvivors. The continuous on-line real-time displays of hemodynamic trends facilitate early recognition of acute circulatory dysfunction.

NMDA Receptor Blockade Prevents Nitroglycerin‐Induced Headaches

Nitroglycerin is commonly used in patients who are at risk for developing cardiac ischemia or infarction, especially in the postoperative period. Patients receiving nitroglycerin often complain of headache. In spite of this side effect, nitroglycerin is still the drug of choice for cardiac ischemia. Nitric oxide, which is liberated from nitroglycerin, appears to play a role in the development of the headache via activation of NMDA receptors. 1-3

Treatment of Resistant Depression in Patients with Cancer with Low Doses of Ketamine and Desipramine

Dear Editor: The incidence of depression in intensive care unit (ICU) patients is high, approximating 50%, with the incidence and severity much higher among terminally ill patients such as those found in a cancer ICU. Many factors can cause depression in these patients, including pain, suffering, a sense of hopelessness, loneliness, and fear of death and the unknown. Those with a history of chronic depression are more susceptible to having a severe recurrence. Many terminal patients who become depressed want to end their life and as a result request euthanasia. In the last 15 years, we have had over 15 patients in our cancer ICU requesting an end to their life. In seriously ill patients the response to current antidepressants, such as selective serotonin reuptake inhibitors (SSRIs), is slow. Additionally, these drugs may be ineffective secondary to poor absorption from the gastrointestinal tract. The patients in our ICU were started on an intravenous combination of low-dose ketamine with fentanyl and concurrently started on low-dose desipramine. The response was rapid, within several hours. These patients had their depression improve despite deterioration in physical status secondary to terminal disease. We have found this combination of medications to have a high success rate with noticeable symptom improvement for the remaining final days of life. Most of the patients require ketamine (6–10 mg=hr), fentanyl (20–30 mg=hr), and desipramine 10–25 mg daily depending on the severity of their disease. Recent papers suggest that ketamine is an effective rapid acting anti-depressant. It is also a useful analgesic without anti-cholinergic side effects. The mechanism of action of ketamine is via blockade of NMDA receptors and glutamate pathways and the possible up-regulation of AMPA receptors. It also decreases the pre-synaptic reuptake of dopamine, norepinephrine and serotonin. Further studies should be undertaken to more fully understand its antidepressant qualities.

Propofol in the treatment of presurgical headache and nausea.

to 1 hour; he treats it with a nonprescription analgesic. The headache is severe in intensity, throbbing over the anterior vertex, and accompanied by photophobia. In October 1996, the patient was watching Clinton’s re-election on TV, which upset him greatly. He subsequently developed his typical visual aura, but it was much more vivid than usual. His visual symptoms were accompanied by tingling of his left arm and hand, developing without a march and lasting for 15 to 30 minutes. He developed a severe headache, photophobia, and generalized weakness. In addition, he was confused and kept repeating the same sentence. The visual disturbance has persisted ever since, affecting his vision as if he is looking through a veil. The “white snakes” in the periphery of both visual fields come and go. Cranial magnetic resonance imaging was normal and without evidence of infarction.

Delayed endotracheal extubation and PTSD in ICU patients.

TO THE EDITOR: Post-traumatic stress disorder (PTSD) is not uncommon in patients in the intensive care unit (ICU). PTSD is often associated with extreme distress, including nightmares, fears, and frightening recollections of the distressful experience. In the literature, ICU-related PTSD was reported to be associated with prolonged mechanical ventilation, sleep disturbances, and the use of narcotics, benzodiazepines, and other sedatives. A recent large multicenter study by Davydow reported that mechanical ventilation for greater than 5 days was a significant predictor of the development of both PTSD and limitations in the patient’s ability to return to usual activities 1 year after the ICU stay. Additionally, this and other studies have shown a pre-existing psychological disorder predisposes patients to developing PTSD in these circumstances. We report our experience as intensivists in a surgical ICU. We have observed a strong correlation between PTSD and patients who remained intubated in the postoperative period, and have noticed two factors in particular that influence this. One is the presence of the endotracheal tube, which caused intense discomfort, pain, and an inability to talk. If the endotracheal tube became blocked with secretions, patients felt a near death experience of choking and asphyxiation. Periodic suctioning with catheters introduced into the trachea was also a very distressful experience. In one extreme case, a patient who was also a physician felt the suction cathe-

Mediastinal haemorrhage mimicking tamponade during en-bloc oesophagectomy.

EDITOR: Concealed mediastinal haemorrhage is rare during oesophagectomy procedures. It may occur in patients with major chest trauma following aortic injury. Occasionally patients undergoing aortic arch surgery may present with postoperative mediastinal bleeding. We report a case in which concealed mediastinal haemorrhage complicated an en-bloc oesophagectomy procedure in the early stages of the operation. The diagnosis was not clear because the bleeding was not obvious, as it involved the posterior mediastinum. The mediastinal haemorrhage mimicked cardiac tamponade and was unresponsive to fluid replacement. A 60-yr-old male, ASA III, with oesophageal cancer, was scheduled to undergo en-bloc oesophagectomy and colon interposition. Past medical history included hypertension, chronic obstructive pulmonary disease, reflux disease and smoking. The surgical plan included initial exploration of the right chest for mobilization of the oesophagus, followed by abdominal and left-sided neck explorations for completion of en-bloc oesophagectomy. The transverse colon would then be mobilized and interposed in place of the oesophagus. Preoperative preparation included blood gas analysis and pulmonary function tests, which were within normal limits. A dobutamine stress echo was negative for cardiac disease. Following induction of anaesthesia, a 39-G double lumen tube was placed uneventfully. Proper placement was confirmed with auscultation and fibreoptic bronchoscopy. Anaesthesia was maintained with isoflurane and intermittent boluses of fentanyl. In addition to the standard anaesthesia monitors, a radial arterial line and pulmonary artery catheter were placed. Initial pulmonary artery pressures were within the normal range (30/15–32/17 mmHg, pulmonary artery occlusion pressures 12–14 mmHg, central venous pressure (CVP) 8–10 mmHg). Initial blood gases were also in the normal range and pulse oximetry showed a saturation of 100% breathing 50% oxygen. The patient was placed in the left lateral decubitus position, one-lung ventilation was initiated and a right thoracotomy was performed in order to mobilize the stomach and oesophagus. This portion of the procedure lasted for the first 3 h of surgery, during which time blood gases remained in the normal range and the patient required only 50% oxygen to maintain adequate oxygenation. After completion of the mobilization, the thoracotomy incision was closed and the patient turned supine. Blood loss at this portion of the procedure was approximately 250–300 mL. Haemodynamic parameters remained in the normal range. Shortly after turning the patient, the patient became haemodynamically unstable, accompanied by a slowly developing fall in oxygen saturation (99–88%). There was no significant change in peak inflation ventilatory pressures. Auscultation of the lungs revealed no abnormalities. The patient was given 100% oxygen. There was minimal drainage from the right-sided chest tube. The blood pressure (BP) ranged from 60/30 to 70/40 mmHg, and heart rate was 120–130 beats min 1. The oxygen saturation continued to deteriorate in spite of 100% oxygen and ventilation of both lungs. Fibreoptic bronchoscopy revealed no abnormal findings. The hypotension was treated with crystalloid and blood. The patient received 3000 mL crystalloid and albumin 5% 500 mL without any significant improvement in haemodynamics. BP was maintained in the 80/40–90/50 mmHg range with 10 μg kg 1min 1 of dopamine. There was no evidence of any bleeding at this time, however, haemoglobin analysis revealed a decrease in haemoglobin from 14.0 to 12.0. The patient received three units of packed red blood cells but there was no response. Pulmonary artery pressures were slightly elevated despite the hypotension. There was no evidence of pericardial tamponade (CVP was 12–15 mmHg, pulmonary capillary wedge pressure was 15–17 mmHg). There was no evidence of ischaemic changes on the electrocardiogram. Cardiac output varied from 4.4 to 5.5 L min 1. The patient continued to require increasing inotropic support and fluids to maintain BP. The surgeon was asked to inspect the mediastinum for any signs of mediastinal bleeding via the upper abdominal incision, as the chest incision was already closed. Inspection of the base of the heart did not show any distension and therefore direct tamponade was excluded. An immediate intraoperative chest X-ray showed widening of the mediastinum with an enlarged upper mediastinal opacity extending to the upper zone of the lung. Initially this opacity was diagnosed as upper lobe collapse. The repeat bronchoscopy showed Correspondence to: Duraiyah Thangathurai, Department of Anesthesiology, LAC/USC Medical Center, USC/Kenneth Norris Cancer Hospital, 1441 Eastlake Avenue, Rm 4341, Los Angeles, CA 90033, USA. E-mail: thangath@usc.edu; Tel: 1 323 865 3441; Fax: 1 323 865 0084

Increased use of protocols in ICU settings

Dear Editor, The use of protocols is becoming more and more common place in American intensive care units (ICUs). The concept of a protocol has its appeal; most physicians feel that certain drugs work best—or at least better than others—in specific situations. Therefore, it seems to make sense to devise a plan of action for those situations in which drugs are utilized in doses based on the patient’s response to them. In this way, almost anyone can simply follow an algorithm to care for the patient, including physicians themselves who, for example, may not be completely familiar with the use of sedatives and/or pain medications in ventilated patients. However, it is our opinion that protocols are currently being overused in ICU settings to the detriment of the patient. Protocols are often necessitated by the fact that in the USA ICU physicians are not always available and are not in-house 24 h a day; therefore, critical care nurses use these guidelines for decision-making. Unfortunately, this results in the simplification of patient care on a very large scale. Each patient will respond differently to each situation, and even drugs routinely used may not have a place in the care of certain patients on specific occasions. Patients may respond better or worse than expected to a specific pharmacological intervention, be it an inotrope, pressor, or sedative. This response needs to be either viewed first-hand by the physician or communicated rapidly to him or her. The appropriate measure may not be to increase the dose of one drug or add the next drug on the list; rather, it may be to change the entire regimen or add a drug not routinely utilized. The blind use of protocols does not allow for this type of intervention and can lead to increased patient morbidity. Even if certain exceptions are built into a protocol, if the protocol is routine enough these exceptions may be ignored. The increased use of protocols also has an impact on the traditional practice of the doctor–patient relationship. The need for hands-on physician interaction with critically ill patients cannot be overstated, as this is the only true means of assessing a patient’s response to an intervention and making appropriate decisions as to whether to increase the dose of a drug, add or delete a drug, or change the entire care plan. While the use of certain medications in a protocol may be a reasonable place to start in the care of a patient, one must have the ability, knowledge, and clinical skills to stray from these guidelines when necessary in order to optimize patient care.

Usefulness of Droperidol as an Anti-Emetic in Terminally Ill Cancer Patients

Dear Editor: Droperidol is a short-acting butyrophenone derivative that has minimal side effects when compared to haloperidol. It is effective as an anti-emetic in doses from 0.625 mg and up through its antidopaminergic properties. It has minimal effects on the respiratory system and has sedative properties that last from 2 to 6 hours. We use droperidol as an adjunct to other opioid drugs, which can induce nausea and vomiting in cancer patients. Droperidol also offers an additional benefit of minimizing psychotic episodes. It has a beneficial effect on ileus that may develop and has minimal anticholinergic properties. It can be used only parenterally, and the action is rapid with intravenous administration. We have been using droperidol for the last 15 years in our intensive care unit for terminally ill patients who suffer from nausea and vomiting. We have found it to be equally as effective as haloperidol, which has anti-emetic properties via the same mechanism as droperidol, but superior to ondansetron, which works through serotonin antagonism (5-HT3 receptor). Droperidol’s sedative and neuroleptic effects make it a useful adjunct for those on very high doses of narcotics. It also has minimal effects on respiration.

Pain relief and sedation in the intensive care unit

p , AIN, fear, and anxiety are common and often overlooked problems in patients in the intensive care unit (ICU). Holland et al 1 have shown that sedative and pain medications are often prescribed to satisfy the expectations of the medical staff or to provide a convenient nursing condition rather than in response to the actual needs and fears of the patient? Pain, fear, and anxiety influence many aspects of the physician-patient relationship. Appropriate relief is often associated with amelioration of the patients complaints even if the underlying disease is not controlled. Historically, even before the advancement of Western medicine, pain and discomfort have concerned every patient and physician and every attempt was taken to alleviate them, frequently without much success. Despite major advances in technology, pharmacology, and pain management techniques, pain in the ICU patient remains largely underevaluated and poorly treated. Discomfort, pain, fear, anxiety, and sleep deprivation result in physiologic and psychological stress with important sequelae. TM Hemodynamic disturbances from catecholamine release may be manifested as hypertension, tachycardia, and cardiac arrhythmias. A neuroendocrine stress response increases production of adrenocorticoid hormones (corticotropin, cortisol, and aldosterone), angiotensin, antidiuretic hormone, and catecholamines, resulting in hyperglycemia, protein catabolism, fluid retention, and electrolyte disturbances. The negative protein balance can lead to poor wound healing and a depressed immune response. The circadian rhythm is reversed, leading to further neuroendocrine and sleep disturbances, s-7 Psychological consequences of pain include insomnia, depression, anxiety, and psychosis. s-t~ It is not surprising that the posttraumatie stress disorder is increasingly reported among patients with the adult respiratory distress syndrome who have been on ventilators. Moreover, patients who are placed on high doses of opiates and benzodiazepines often subsequently develop acute withdrawal states. PROBLEM AREAS IN CRITICALLY ILL PATIENTS