The Drugs Don’t Work: (Non)management Of Postoperative Nausea And Vomiting
Author: Dr Karen Stuart-Smith, Consultant Anaesthetist, Glan Clwyd Hospital, Denbighshire, Wales

Ultrasound Guidance: A Cost-effective Approach To Improving The Success Of Regional Anaesthesia
Author: Tracey Byard

Slaying The POSSUM. Improving Surgical Scoring Systems To Improve Patient Care
Author: Dr Alexandra Ford. Specialist registrar in Anaesthesia and Intensive Care, Glan Clwyd Hospital, Denbighshire, North Wales

The Drugs Don’t Work: (Non)management Of Postoperative Nausea And Vomiting
Author: Dr Karen Stuart-Smith, Consultant Anaesthetist, Glan Clwyd Hospital, Denbighshire, Wales

Recently, an anaesthetic specialist registrar working with me did an audit of the incidence of postoperative nausea and vomiting in our hospital. Patient numbers were very small, and by the standards of a proper clinical trial, it was so underpowered it could not have moved a child’s bicycle. In spite of this, it showed a worryingly high incidence of postoperative nausea and vomiting (PONV) in the face of the apparent appropriate administration of antiemetic therapy, particularly in the ward setting. This prompted me to read the literature on PONV, to determine whether the work of others could shed light on our clinical problem. Several facts became obvious. These are: 1) the mechanism of any form of nausea and emesis is still not at all understood, 2) very little new research is being conducted in this area, and 3) none of the various treatments, including the very newest, are any better than those that have been available for 30 years. Although any antiemetic is (nearly) better than no antiemetic in a high-risk patient, identification of the high risk patient remains problematic. On the other hand, treating everybody just in case is probably not cost effective, and runs the risk of side effects (although what the clinical impact of these side effects might be, no-one knows1). On a possibly more positive note, it struck me that the small sample population contained in our small local audit did in fact exhibit the characteristics predicted by the current literature. This article will discuss our current understanding of PONV (or lack of it), and then I will indulge myself by describing our audit and how the patients we studied fitted into the current pattern.

The physiology of nausea and vomiting is still unclear. One of the principle problems with research in this area is that very few animals experience vomiting: dogs, ferrets and shrews are the main animal models. Only humans can express a sensation of nausea, which makes patients the only available experimental model for this symptom. Nausea and vomiting are probably mediated via different neural pathways. For example, in both PONV and chemotherapy-associated emesis, ondansetron, a 5-HT3 receptor antagonist, is more effective at preventing vomiting, whereas droperidol, a dopamine (D2) receptor antagonist, is effective against nausea^2,3. In fact, a depressing multitude of brainstem receptors (in the area postrema and nucleus tractus solitarius) appear to be involved in nausea and vomiting4, and the gut, the major source of 5-HT, contains several different types of 5-HT receptor which may be of relevance to both PONV and chemotherapy-induced emesis (for review see reference 5). In addition there are probably significant gender differences in the response to antiemetics: not only are females at significantly higher risk of PONV⁶, but the therapeutic effect of droperidol may be limited to this sex⁷. In other words, the perceived increased incidence of PONV in laparascopic surgery is more likely to be related to the fact that laparascopic procedures such as cholecystectomy and (obviously) gynaecological surgery are more often performed in women rather than men, i.e. the sex of the patient may be a more important risk factor for PONV than the type of surgery⁷. However type of surgery has a more complex bearing on the development of PONV because the use of opioids for pain relief, and the use of volatile anaesthetics, particularly over a long period of time, strongly influence the development of PONV⁸. With regard to the induction and maintenance of anaesthesia, it is worth noting that although propofol is associated with a reduced risk of PONV, the apparent protective effect of propofol occurs only when it is used as part of a TIVA technique for maintenance of anaesthesia. Propofol given as a single shot for induction has no anti-emetic effect. Propofol TIVA may inhibit 5-HT3 receptors as part of its anti-emetic action. However many workers contend that it is the avoidance of inhalational agents rather than the use of propofol per se which accounts for its effect.

A further complication is the phenomenon of early versus late PONV. The exact definition of ‘early’ versus ‘late’ symptoms has not been generally agreed, but for the purposes of this article, I will offer a personal definition of ‘early’ PONV as occurring within 6 hours of surgery, and more specifically in the recovery room. ‘Late’ PONV will be considered to be any time beyond the first 6 hours, and might even include two separate time periods: ‘late’ referring to the phase 6-24 hours after surgery, and ‘very late’ 24-48 hours after surgery. My rationale for these timings will (I hope) become clear during the course of this article. It is worth noting that early and late symptoms of nausea and vomiting are also recognised in cancer chemotherapy: acute emesis occurs within 24 hours of commencing chemotherapy, and late emesis after the first 24 hours⁹. 5-HT3 antagonists prevent the early but not the late emesis, and have very limited efficacy for nausea at any stage in this group of patients^3,9. There are interesting parallels with PONV, where ondansetron is recognised to be better at preventing vomiting than nausea¹⁰.

Early PONV is relatively uncommon, at least in the recovery room. This is probably because intraoperative antiemetics are given as a matter of routine. It has been suggested that PONV is almost uneconomic to treat1, except in high risk groups¹¹, especially as blanket treatment of all patients may lead to an unacceptable rate of side effects¹. Nevertheless, PONV is a source of considerable distress to patients. Surveys have shown that patients would be willing to fork out considerable sums of their own money if they could buy their way out of PONV¹⁰. The key is to determine who the high risk patients are. The classical four major risk factors, female gender, previous history of motion sickness or PONV, being a non-smoker, and the use of opiate analgesics, are cumulative, so that the more risk factors that are present in any one patient the more likely they are to suffer PONV¹¹. The problem is where does one draw the line? Should all females receive anti-emetics, or only the non-smokers? Is a man with a history of motion sickness, who is also a smoker, likely to experience PONV or not? We simply do not know enough about the basic science of each precipitating factor to make an informed decision.

Early PONV is taken very seriously, and a great deal is written on the subject. By contrast, late and very late PONV is not taken at all seriously, even though it is a very common phenomenon. I would suggest that as anaesthetists, we consider a good result to be a lack of nausea in the recovery room, but fail to acknowledge or take responsibility for any nausea and vomiting that takes place after the patient has been wheeled back to the ward and/or sent home. Late PONV is an anaesthetic problem, if only because precipitating factors, such as pain, opioid administration and dehydration, are within the sphere of influence of the anaesthetist even after the patient has disappeared into the lift and thus out of our sight. Very late PONV, i.e. nausea and vomiting experienced after the patient has gone home, is far more common than is generally realised. A fascinating study showed that 35% of all patients suffered PONV for up to 5 days (the time limit examined in the study) after discharge, with symptoms in the first 24-48 hours being the most common¹². These symptoms interfered significantly with activities of daily living: eating, preparing meals, running errands, performing household chores, taking care of themselves and others. None of these patients had been given pre-discharge advice on the treatment of nausea and vomiting (although a significant number of them probably went home armed with analgesics such as co-codamol, tramadol, etc). Of particular interest is that the vast majority of these patients (about 72%) had not experienced any early (recovery room) PONV. Thus, the cheerful day surgery patient who you anaesthetised for that laparoscopic sterilisation may look rewardingly nausea-free in recovery, but may be vomiting rings round herself for two days when she gets home, and is having difficulty caring for her family.

Should we accept at least some of the responsibility for this? Of course we should. At the very least, we should be examining the reasons for this later onset PONV, and designing strategies to manage it. The difficulty lies in part in the lack of suitable post-discharge anti-emetics and adequate post-discharge care. The problem is that the average Joe or Jill sent home after day surgery has only an over-worked district nurse and a poorly accessible GP appointment system to help him or her out. There are no reliable over-the-counter remedies. Metoclopramide is frequently prescribed by GPs, but the true anti-emetic dose is probably in the region of 50mg rather than the generally administered 10mg. Such a dose carries a high risk of drowsiness and extrapyramidal symptoms, and cannot be recommended for general use¹². Ondansetron melts are another possibility, but there is considerable debate over its efficacy in late emesis. Ondansetron is of limited usefulness in delayed nausea and vomiting after chemotherapy9, and the same may be true of late PONV, although there are no studies available to determine this. Ondansetron may also be more effective given prophylactically rather than after development of symptoms, and as described above, 5-HT3 antagonists in general are better at preventing vomiting than nausea. Higher doses incur the risk of a headache. This raises the scenario of postoperative patients being sent home with a fistful of ondansetron, which they have to take regularly to prevent the onset of vomiting. In addition they may still feel nauseous, tempting at least some into taking larger doses of the drug, with the potential of a headache adding to their woes. In patients undergoing cancer chemotherapy, ondansetron is usually balanced with dexamethasone, which is more effective at eliminating nausea. Dexamethasone has some potential merit in late and very late PONV. At the very least it is an effective adjunct to 5-HT3 antagonists6. Of greater importance is that dexamethasone is particularly effective in late-onset nausea and vomiting¹³. I would suggest that low-dose dexamethasone is worth investigating as a therapy for late PONV. Of course concerns must be raised about the possibility of inhibition of the hypothalamic-pituitary-axis. There are no studies available on this topic. However it is unlikely that very low-dose dexamethasone administered for a few days will have a significant effect in a healthy patient, and there has been no evidence of inhibition of intrinsic steroid production in cancer patients receiving short-term, high-dose dexamethasone therapy for chemotherapy-related emesis⁸.

Before leaving the general topic of PONV therapy, some words must be said about other anti-emetic therapies, old and new. I particularly want to discuss droperidol, previously known to be a very effective anti-emetic, but withdrawn globally after the American Food and Drugs Administration issued a ‘black-box’ warning about this agent in 2001. The danger was thought to be a droperidol-induced prolongation of the QT interval (long QT syndrome, LQTS) on the ECG, which could progress to torsade de pointes (TdP). This conclusion was based on a small number of case reports gathered over 30 years¹⁴. For the uninitiated, TdP is a paroxysmal ventricular tachycardia, characterised by varying amplitude of the QRS complex and rotation of the QRS axis around the isoelectric point. It is generally self-limiting, but can deteriorate into ventricular fibrillation and thus is potentially fatal. LQTS can be congenital (as a result of genetic alterations in cardiac potassium channels), but it is also precipitated by acute myocardial infarction, electrolyte disturbances and hypothyroidism. Antipsychotic agents in the haloperidol group are classically associated with QT-prolongation, and the most widely-quoted (especially in medical multiple-choice exams!). However many other drugs, including antifungals, erythromycin, various anti-arrhythmics (including sotalol, a widely prescribed cardiac drug), tricyclic antidepressants, and indeed a wide variety of agents, alter cardiac ion channels and predispose to LQTS. None of these are banned by the FDA. Particularly puzzling is that 5-HT3-receptor antagonists also predispose to QT-prolongation¹⁴, but again, a black box warning has never been contemplated. In a thoughtful assessment of QT-prolongation and psychiatric agents, actually published in the same year as the black-box warning on droperidol¹⁵, it was concluded that although both haloperidol and droperidol can cause this ECG anomaly, it is hard to pin down more than a tiny handful of cases where haloperidol might have been the precipitating factor. This study showed a much clearer association between thioridazine and fatal TdP, although the incidence was still low¹⁵. Needless to say, thioridazine is still also in circulation, unlike droperidol. The blame-game with droperidol is even harder to justify when it is remembered that QT-prolongation is a sign of potential problems, not a definite predictor of the onset of TdP. In general, the degree of prolongation of the QT-interval caused by a particular drug is not a guide as to whether it will induce a ventricular dysrrhythmia^14,15. For all these reasons, the targeting of droperidol by the FDA is unjustified. A very recent analysis of droperidol usage and the occurrence of TdP in one centre in the 4-year period leading up to the black-box warning concluded that during this time, approximately 17000 patients received droperidol as prophylaxis for PONV, but there were no recorded instances of TdP. The conclusion of this study was that the FDA black box warning for droperidol is ‘excessive and unnecessary’.

Does this matter? The answer is it might. Quite apart from the general ethical implications of ill-informed government intervention in drug therapy (NICE take note), droperidol was a simple, cheap and relatively effective anti-emetic⁷. A maximum dose of 1.25 mg is required for efficacy. The most recent guidelines on the subject, by the Society for Ambulatory Anaesthesia, stress that this would have been that group’s anti-emetic of choice had the black box warning not been in place¹⁶. Although I probably wouldn’t go that far, droperidol does have some proven benefit when combined with PCA, and is also a useful adjunct to ondansetron¹⁷, especially as it may have a protective effect against ondansetron-induced headache.

Finally in this section, some brief words about the newest antiemetic about to come on the market. Neurokinin NK1-receptor antagonists, such as aprepitant, inhibit the action of substance P at various sites in the area postrema and nucleus tractus solitarius⁹. Substance P is released from the enterochromaffin cells of the gut, and acts on the brainstem via NK1-receptors to induce emesis. Results in chemotherapy have been mixed, with an improved efficacy in comparison to ondansetron in the setting of some chemotherapeutic agents but not others³. In PONV, aprepitant is of equal potency with ondansetron for early symptoms, but may be of greater benefit in late nausea and vomiting, which would be a significant advantage¹⁸. It is worth noting, however, that aprepitant induces increased warfarin metabolism through an effect on cytochrome P450 enzymes, and reduces the efficacy of the oral contraceptive pill³. Although these factors are relevant for long-term antiemetic therapy in cancer patients, it is too early to assess the relevance of these drug interactions in PONV.

Getting back to the topic of the first paragraph in this article, where does all this information on PONV leave the patients in our own modest little audit¹⁹? The audit was conducted by Dr Sonia Flory, a Specialist Registrar in Anaesthesia at Glan Clwyd Hospital. We looked at a random collection of patients passing through the recovery room each day. All were undergoing elective surgery. Preoperative risk factors, intraoperative management, and postoperative factors were considered in relation to the development of very early (in recovery room) and later (on the ward and at home) PONV. Information was only gathered on a total of 89 patients (reflecting the difficulties our juniors face in getting audits done in the current NHS environment). Nevertheless a history of previous PONV was a strong risk factor for the development of PONV on this admission, reflecting the importance of this risk factor in the population in general. In this group, receiving an antiemetic did not reduce their incidence of PONV. However very few patients in this high-risk group received more than one antiemetic. Most guidelines would recommend a combination of antiemetics in those with a previous history of PONV-dexamethasone and ondansetron, for example, although other combinations are possible. This group aside, the striking finding of the audit was that while the overall incidence of PONV in the recovery room was very low (12.4%), the incidence of PONV amongst patients undergoing major general surgery was relatively high: 43.8% of all recovery room PONV occurred in this group. Potential risk factors probably include prolonged use of volatile anaesthetic agents and (perhaps) the use of intraoperative opioids, although overall across the whole study group opioid use was not a factor in this early PONV. Total intraoperative opioid dose might be the important component here, but we did not look at this factor (in general total opioid dose has never been specifically looked at as an independent risk factor in PONV in the literature). A striking feature was that nearly half of the older patients in this study (age > 65 years), did not receive an intraoperative antiemetic. Although it is difficult to prove with our small numbers, it is likely that this group accounted for a significant amount of the early PONV seen in the major general surgical patients. Age is not generally considered a risk factor for PONV, possibly because many studies evaluating new antiemetic therapies have focussed on a handy patient group that is healthy enough to be included in clinical trials but exhibit a satisfying degree of nausea and vomiting: young women having laparascopic surgery. I think this may have led to the unwarranted assumption (and I include myself in this) that PONV is much more a feature of young women rather than old men. In fact anyone receiving prolonged surgery (increased stress response), using volatile anaesthetics and opioid analgesia, is probably at increased risk of PONV, and combination antiemetic therapy is probably warranted in this group. The results of this audit have certainly changed my practice in this regard.

What about late PONV? 53 of the 89 patients in our audit were in-patients (i.e. not day surgery). This group had a ward-based PONV rate of 43.4%, and most of these patients were over 65, and had had major general surgery. In other words, late (ward) PONV was a bigger and more significant problem than recovery room PONV, which is in line with the discussions above. Significant rates of PONV were detected up to 48 hours post-operatively, again in agreement with published data. In almost all cases, antiemetics were given to patients who experienced nausea or vomiting on the ward, so PONV was detected and treated. However the presence of symptoms was the trigger for instituting therapy, and the literature tends to suggest that prophylaxis is much more effective than treatment, especially for ondansetron. There is no clear information for cyclizine. Dexamethasone may be an effective rescue antiemetic in the presence of established PONV, but the literature is hazy on all of this so I have not included individual references. It is certainly something that is worthy of a proper randomised clinical trial. The precipitating cause for the high rate of ward PONV is not clear, although high pain scores correlated with an increased incidence of nausea. Our study was not designed to determine whether the nausea was caused by the pain or by the opioid used to treat it. However one might suspect both PCA and tramadol were implicated here. The most recent guidelines on PONV glibly suggest that regional analgesia is an effective means of avoiding PONV¹⁶, although there are no clinical trials available to support this contention. Again, this is another area ripe for proper multicentre research. It is also worth remembering that intrathecal opioids can induce emesis, and both epidural and intrathecal opiates can cause unpleasant itching.

We did not have sufficient returns to assess the level of ‘at home’ PONV in day surgery patients. However given the results of the Carroll study described above¹², and given that we are strongly encouraged by politicians to discharge our patients earlier and earlier (the Americans call this phenomenon ‘quicker and sicker’), we will be looking much more closely at this aspect in the future. The Carroll study was published in 1995, and is probably ripe for updating.

Although many more clinical trials remain to be done in PONV, they need to be more targeted to individual clinical factors rather than repeated ‘catch all’ multicentre trials. One good example is the separation between nausea and vomiting. Based on our own observations, nausea is by far the more common symptom, but it causes a great deal of patient distress. It is simply not treated well. We need to be as rigorous about treating PONV on the ward as we are about fluid management and pain relief. The principle problem though is the paucity of sensible basic science in the area. I searched the literature in considerable detail, but could find almost nothing new on the central mechanisms of nausea and vomiting since about 1990. Most new work on gut physiology is being done by physicians on relatively uncommon gut motility disturbances, while the very common problem of PONV has been largely neglected. The huge variety of brainstem receptors with the potential to mediate different aspects of the nausea and vomiting process has led to a random, scattergun approach to management, of the ‘here’s a receptor let’s block it’ mentality. Although newer drugs such as aprepitant show promise, there is not yet convincing evidence that they make enough of a difference to merit a hole in the hospital budget to pay for it. All the available evidence shows that on the whole, the drugs don’t work in PONV. It is time for a new approach and a fresh paradigm.

References

1. Carlisle JB and Stevenson CA. Drugs for preventing postoperative nausea and vomiting. Cochrane Database of Systematic Reviews 2006. Issue 3. Art. No. CD004125
2. Chan MTV et al. The additive interactions between ondansetron and droperidol for preventing postoperative nausea and vomiting. Anesth Analg 2006;103:1155-62
3. Warr DG. Chemotherapy- and cancer-related nausea and vomiting. Current Oncology 2008;15 (Suppl 1):S4-9
4. Chepyala P and Olden KW. Nausea and vomiting. Current Treatment Options in Gastroenterology 2008;11:135-44
5. De Ponti F. Pharmacology of serotonin: what a clinician should know. Gut 2004;53:1520-35
6. Kim EJ et al. Combination of antiemetics for the prevention of postoperative nausea and vomiting in high-risk patients. J Korean Med Sci 2007;22:878-82
7. Apfel CC et al. A factorial trial of six interventions for the prevention of postoperative nausea and vomiting. N Engl J Med 2004;350:2441-51
8. Ho KY and Chiu JW. Multimodal antiemetic therapy and emetic risk profiling. Ann Acad Med Singapore 2005;34:196-205
9. Saito R et al. Roles of substance P and NK1 receptor in the brainstem in the development of emesis. J Pharmacol Sci 2003;91:87-94
10. Habib AS and Gan TJ. Evidence-based management of postoperative nausea and vomiting: a review. Can J Anesth 2004;51:326-41
11. Beattie WS. Strategies to reduce postoperative nausea and vomiting: does metoclopramide have a role? Can J Anesth 2002:49:1009-15
12. Carroll NV. Postoperative nausea and vomiting after discharge from outpatient surgery centres. Anesth Analg 1995;80:903-9
13. Henzi I et al. Dexamethasone for the prevention of postoperative nausea and vomiting: a quantitative systematic review. Anesth Analg 2000;90:186-94
14. Nuttall GA et al. Does low-dose droperidol administration increase the risk of drug-induced QT prolongation and torsade de pointes in the general surgical population? Anesthesiology 2007;107:531-6
15. Glassman A and Bigger JJ. Antipsychotic drugs; prolonged QTc interval, torsade de pointes, and sudden death. Am J Psychiatry 2001;158:1774-82
16. Gan TJ et al. Society for Ambulatory Anaesthesia guidelines for the management of postoperative nausea and vomiting. Anesth Analg 2007;105:1615-28
17. Chan MTV et al. The additive interactions between ondansetron and droperidol for preventing postoperative nausea and vomiting. Anesth Analg 2006;103:1155-62
18. Diemunsch P et al. Single-dose aprepitant vs ondansetron for the prevention of postoperative nausea and vomiting: a randomized, double-blind phase III trial in patients undergoing open abdominal surgery. Br J Anaesth 2007;99:202-11
19. Flory S and Stuart-Smith K. Age discrimination in the management of postoperative nausea and vomiting. Poster presentation to the Society of Anaesthetists of Wales, March 2008.

Ultrasound Guidance: A Cost-effective Approach To Improving The Success Of Regional Anaesthesia
Author: Tracey Byard

Peripheral nerve blockade has traditionally relied on anatomical landmarks and electrical stimulation to localise nerves, but this approach can be inaccurate and the resulting misplacement of the needle can cause blocks to fail. The technical demand of regional blocks means that, depending on the technique used, failure rates range from 5-25%¹. Multiple trial-and-error attempts to place a needle can cause unwarranted pain to the patient, waste valuable time in theatre and, in some cases, lead to potentially serious complications.

Although many complications could be anticipated and prevented by good management of the predictable physiological side effects to the anaesthetic, direct nerve injury can sometimes occur during peripheral nerve blockade guided only by anatomical landmarks and nerve stimulation. The implications for the patient are serious, particularly because very little effective therapeutic interventions are available to treat the complications, with permanent loss of function as the likely outcome. It is therefore important to focus on the prevention of these serious complications which are potentially avoidable.

The popliteal nerve, pre-, during and post-injection, as clearly shown by the SonoSite MicroMaxx hand-carried ultrasound system

Ultrasound in anaesthesia – clear advantages
Anaesthetists first started to use ultrasound imaging to guide nerve blocks in surgical patients in the mid-1990s^2,3, providing non-invasive real-time visualisation of the target nerve and surrounding vasculature, as well as the needle and spread of injection. More recently, evidence has been accumulating proving the great potential this technique offers to reduce the number of failed blocks and risk of complications, as well as improving the efficacy of successful blocks. Schwemmer et al. (2006)⁴ directly compared ultrasound guidance and neurostimulation in axillary brachial plexus blocks, and found that 19.7 % of patients treated by nerve stimulation required an additional dose of opioid or general anaesthesia due to incomplete or insufficient block, compared to just 3.6 % of patients treated by ultrasound guidance. Furthermore, only 5.4 % of patients in the ultrasound group required monitoring in the post-anaesthesia care unit, whereas 32.4 % of the neurostimulation group required postoperative monitoring.

Successive studies have demonstrated many advantages of ultrasound guidance when compared with nerve stimulation for performing a variety of nerve block types in adults and children, including the faster speed of onset of anaesthesia^4,5,6, improved quality or duration of the block^3,5,7, and reduction in the mean required volume of anaesthetic^7,8.

The cost-effective option
The initial investment required for an ultrasound system can vary hugely, but the technology is advancing rapidly and hand-carried systems are now available at a significantly lower cost than large cart-based units. Importantly, the imaging quality and resolution of modern hand-carried systems are comparable to those of the more expensive cart-based systems. Some systems, such as the SonoSite S-Nerve™ ultrasound tool, have been specifically designed for use in regional anaesthesia procedures and are equipped with software that enables clear visualization of nerves. SonoSite has developed easy to use hand carried ultrasound systems that allow the user to focus on the procedure, not the ultrasound.

Two recent studies have looked closely at the overall costs involved in performing nerve blocks using either nerve stimulation or ultrasound guidance. Sandhu et al. (2004)⁶ compared the costs of infraclavicular blocks administered by nerve stimulator and ultrasound guided techniques using a portable ultrasound system. They calculated that the cost of the hand-carried system with the relevant probe (SonoSite™ 180PLUS with C11 transducer) was approximately £8,370 (€12,410) which, based on performing 5,000 blocks, equates to an average cost of £1.67 (€2.48) per block. However, this initial outlay does not take other costs into consideration where using ultrasound was found to save considerably, including savings per patient of £0.59 (€0.88) for equipment; £19.70 (€29.20) for operating room time; and £63.00 (€93.50) for the latency of a block. These figures come to an estimated saving of £83.29 (€123.58) per patient when ultrasound was used, with savings of £82.70 (€122.70) per patient due to time savings alone. The published data showed that the investment made in a SonoSite hand-carried ultrasound system could be recouped in as few as 100 patients.

Schwemmer et al. (2006)⁴ came to similar conclusions when they compared neurostimulation and ultrasound guidance methods in axillary brachial plexus blocks. They found that the speed of anaesthetic onset was significantly faster (20 minutes) when blocks were performed using ultrasound, and calculated that each patient who received ultrasound guidance saved the hospital an average of £78.80 (€116.80), based on the previously published calculation¹¹ that one minute of surgical time costs £3.90 (€5.80). The study also found that using nerve stimulation required an extra £35.60 (€52.80) in total staff costs and £6.70 (€10) in general anaesthesia costs.

Training
Training is an important issue for ultrasound guidance as the procedure requires the ability to recognise nerves and other relevant anatomical structures in images. There are a number of short ultrasound workshops and courses now widely available that are aimed specifically at non-radiologists. Interestingly, one study⁴ noted that when patients received brachial plexus blocks guided by a hand-carried ultrasound system, the time required to establish the block decreased as the operator’s experience with ultrasound increased, clearly demonstrating a rapid learning curve.

Contiplex and Stimuplex are registered trademarks of B Braun Melsungen AG. FlexTip Plus is a registered trademark of Arrow International, Inc. Exchange rates as at 31 July 2007, $1=£0.49232; $1=€0.73012

Table 1: Cost savings of ultrasound-guided technique over using a nerve stimulator for regional blocks (table based on Sandhu et al. 2004). The cost of operating room time was calculated at £3.90 (€5.80) per minute. The data demonstrate that a large proportion of the savings come from time spent on injection and time to onset of anaesthesia.

Summary
Ultrasound guidance allows physicians to directly visualise relevant nerve structures for upper and lower extremity nerve blocks at all levels in real-time, significantly improving the quality and outcome of most peripheral nerve blocks. The technique brings considerable time and cost savings to hospitals, as well as improving safety and reducing pain for the patients, and a NICE interventional procedure guideline is anticipated for ultrasound-guided regional anaesthesia.

Clearly there is significant evidence that ultrasound is indeed a cost-effective option for guiding regional anaesthesia. As well as offering convenience and portability, hand-carried ultrasound systems save time and money when compared to cart-based systems. For example, systems that are not PC-based (such as the SonoSite S-Nerve™ ultrasound tool) are ready to use within seconds of switching the machine on, saving considerable amounts of time. Many physicians have been disillusioned by attempting to use complicated ultrasound systems borrowed from radiology departments but some systems available now are specifically designed for regional anaesthesia, and are consequently much more easy to use for guiding nerve blocks.

This article is based on another article entitled Improving Regional Anaesthesia featured in the November 2007 issue of The Clinical Services Journal.

References

1. Fischer B (2007). Complications of regional anaesthesia. Anaesthesia and Intensive Care Medicine 8: 151-154
2. Kapral S, Krafft P, Eibenberger K, Fitzgerald R, Gosch M, Weinstabl C (1994). Ultrasound-guided supraclavicular approach for regional anaesthesia of the brachial plexus. Anesth Analg 78: 507-513
3. Marhofer P, Schroegendorfer K, Koinig H, Kapral S, Weinstabl C, Mayer N (1997). Ultrasonographic guidance improves sensory block and onset time of three-in-one blocks. Anesth Analg 85: 854-857
4. Schwemmer U, Schleppers A, Markus C, Kredel M, Kirschner S, Roewer N (2006). Process management in axillary brachial plexus block. Anaesthesist 55: 451-456
5. Williams SR, Chouinard P, Arcand G, Harris P, Ruel M, Boudreault D, Girard F (2003). Ultrasound guidance speeds execution and improves the quality of supraclavicular block. Anesth Analg 97: 1518-1523
6. Sandhu NS, Sidhu DS, Capan LM (2004). The cost comparison of infraclavicular brachial plexus block by nerve stimulator and ultrasound guidance. Anesth Analg 98: 264-269
7. Oberndorfer U, Marhofer P, Boesenberg A, Willschke H, Felfernig M, Weintraud M, Kapral S, Kettner SC (2007). Ultrasonic guidance for sciatic and femoral nerve blocks in children. Br J Anaesth 98: 797-801
8. Casati A, Baciarello M, Cianni SD, Danelli G, De Marco G, Leone S, Rossi M, Fanelli G (2007). Effects of ultrasound guidance on the minimum effective anaesthetic lume required to block the femoral nerve. Br J Anaesth 98: 823-827

Slaying The POSSUM. Improving Surgical Scoring Systems To Improve Patient Care
Author: Dr Alexandra Ford. Specialist registrar in Anaesthesia and Intensive Care, Glan Clwyd Hospital, Denbighshire, North Wales

Introduction
Scoring systems have existed for many years as tools for assessing patient morbidity and mortality and can be used to influence treatment decisions. They may also be used to stratify patients into groups according to the severity of their illness so that comparisons of outcome can be made between different units.

Scoring systems may be specific to a particular disease (e.g. Ransons’ criteria in pancreatitis), or system (e.g. Goldman Cardiac Index). They may be based on chronic health or on current physiological parameters.

This article aims to discuss the development of these scoring systems and their relevance in today’s practice, with particular reference to their usefulness in predicting outcome in high-risk patients with peritonitis requiring major laparotomy.

An overview of physiological scoring systems
A number of scoring systems have been developed over the years which aim to determine the severity of illness based on the physiological response to it. The most well known and commonly used of these is the Acute Physiology and Chronic Health Evaluation score, best known as APACHE.

The original APACHE score used 34 physiological variables which were obtained in the first 24 hours of admission to ICU along with information on the chronic health of the patient. Its use was limited by the large amount of information required to generate a score.¹ In 1981 this was modified by Knaus et al to the APACHE 2 score which used only 12 variables. This improved compliance and was still found to be useful in predicting outcome in critically ill patients and thus became the most widely used of the risk scores.^{2, 3} An APACHE 3 score has also been developed with a predictive performance that is comparable to, or perhaps even superior to APACHE 2 however it is not commonly used. This is because the equations it employs are commercially protected.^{4, 5}

Other physiological scores were developed and investigated. The Simplified Acute Physiological Score 2 (SAPS 2) and Sickness Severity Score (SSS) are both modifications of APACHE. SAPS 2 used 13 physiological variables and also took into account the type of admission (elective/emergency; medical/surgical). The Sickness Score used general status rather than worst values to allow for transient changes. Both of these were found to be comparable to APACHE 2 in prediction of outcome but neither gained the same popularity.³

An overview of scoring systems for intra-abdominal pathology and peritonitis
So far the systems discussed have been generally used for critically ill patients. The cause of the critical illness has not been regarded and the patients may be medical or surgical. For the general anaesthetist, however, an accurate estimate of morbidity and mortality in the acute surgical patient presenting for emergency procedures is of much greater relevance. Of particular interest are the scoring systems specifically developed to predict outcome in critically ill patients with intra-abdominal contamination. These patients may have grossly deranged physiology and may also require major surgical intervention so standard scoring systems may not adequately predict morbidity and mortality.

A multitude of scoring systems for acutely ill surgical patients have been developed. Some, such as the Multi Organ Failure score (MOF), Organ System Failure score (OSF) and Acute Organ System Failure (AOSF) score, use the number of failing organs and the degree of organ failure to determine outcome. However in patients with intra-abdominal pathology only the OSF score was found to compare with APACHE 2 in predicting mortality.^{3, 6, 7, 8, 9}

The Mannheim Peritonitis Index (MPI) uses information on age, sex, organ failure, malignancy and details of the peritoneal contamination to develop a score to predict mortality which is comparable to APACHE 2.^{10, 11, 12 14} The Biochemistry and Haematology Outcome model (BHOM) and Surgical Risk Score (SRS) were also found to be able to predict outcome after emergency surgery.^{13, 15} Other scores such as the Sepsis Severity Score and the Peritonitis Index Altona when assessed were not able to predict outcome with the same accuracy in intra-abdominal sepsis.^{10, 16}

Despite the interest in developing new risk scoring tools, the APACHE 2 score remained the most widely used in practice until the POSSUM was introduced.

Development of Colorectal POSSUM
The main disadvantage of APACHE 2 from a surgical point of view is that it does not consider the operation that the patient is due to undergo. It is apparent that the predicted outcome of a patient due to undergo the incision and drainage of an abscess would be very different from a physiologically similar patient due to undergo a major laparotomy. A system was required that took into account both the physiological status of the patient and the significance of the surgical procedure and thus, the Physiological and Operative Severity Score for the enumeration of Mortality and morbidity (POSSUM) was developed.^{17, 18}

The POSSUM data set was developed over 2 years by Copeland et al. Risk factors for adverse surgical outcomes were assessed by multivariate analysis to detect which predicted complications and mortality. Of the 35 initial risk factors, 12 were detected to be independently predictive of outcome and were included in the physiological data set. These variables were then graded and scored exponentially as 1, 2, 4 or 8 with 1 being in the normal range and 8 being the most deranged physiological value. (See table 1)17

Table 117

Physiological Parameters

Age

Cardiac History

Respiratory History

Blood Pressure

Pulse Rate

Glasgow Coma Score

Haemoglobin

White cell count

Urea

Sodium

Potassium

Electrocardiogram

The other data set in the POSSUM score was operative severity which contained 6 operative variables also based on an exponential score from 1 to 8 similar to the physiological score. (Table 2) ¹⁷

Table 217

Operative Parameters

Operative severity

Multiple procedures

Total blood loss

Peritoneal soiling

Presence of malignancy

Mode of surgery

The two scores were combined and underwent logistic regression analysis to generate a risk equation which gives a percentage score of morbidity and mortality.

POSSUM scoring was evaluated extensively and in general was found to usefully compare outcomes between hospitals and surgeons if used correctly.^{19, 20}

POSSUM scoring was quickly embraced by the surgical community. However, a group in Portsmouth evaluated the POSSUM against their mortality rates and found that it grossly misrepresented their outcomes. They discovered in their cohort of patients that the POSSUM equation over-predicted death by more than a factor of 2. This was even more significant in low risk groups where it over predicted death by a factor of 6! They discovered that by using the original equation the lowest achievable score predicted a mortality rate of over 1%. This figure conflicts with other audits which suggest that low risk patients undergoing low risk surgery have a mortality that approaches 0%.^{19, 21, 23}

These authors used their sample data to develop their own Portsmouth predictor equation (p-POSSUM) using logistic regression. The physiological and operative severity data set was kept as the original POSSUM. The new equation gave a minimal risk of mortality as 0.2% which is more believable but still probably too high.^{19, 21, 22, 27}

In some studies P-POSSUM was found to be superior to the original POSSUM for predicting outcome in low risk groups however, both systems seemed to grossly underestimate mortality in elderly, high risk patients undergoing emergency surgery. (20) It also became apparent that they were still limited when applied to patients undergoing colorectal, upper gastrointestinal or vascular surgery. Thus, specialty-specific POSSUM scoring began to be developed which allowed for case-mix to be accounted for in complex, high–risk procedures such as vascular (v-POSSUM), oesophagogastric surgery (o-POSSUM) and of specific interest to this article, colorectal POSSUM (CR- POSSUM).²⁴

Table 324
Physiological Score	Operative Score
Age group (years)	Operative severity
Cardiac failure	Peritoneal soiling
Systolic blood pressure	Operative urgency
Pulse rate	Cancer staging
Urea
Haemoglobin

CR-POSSUM was developed between 1993 and 2001 on data collected from 6883 patients in 15 hospitals. The information was taken from all patients undergoing emergency or elective colorectal surgery for curative, palliative or diagnostic reasons and observed and expected

mortality rates were compared. The CR-POSSUM data set uses fewer variables (6 physiological variables and 4 surgical variables; see table 3) and is thus easier to use. Most, but not all, studies have found it to be a more accurate predictor of in-hospital mortality for colorectal cases than p-POSSUM.^{24, 25, 26}

Limitations of CR-POSSUM in patients with peritonitis
Despite being generally a better predictor of mortality in surgical patients than other scoring systems it still has its inadequacies particularly when considering patients with intra-abdominal contamination. These patients have grossly deranged physiology secondary to the disease process which have a profound influence on the success of the operative procedure. This includes the development of the systemic inflammatory response syndrome (SIRS) with the resultant implications for oxygen delivery, coagulation, and renal function. It is unlikely that the reduced number of physiological parameters, particularly the low number of biochemical markers, used in CR-POSSUM is reflective of the complexity of the disease. Only haemoglobin and urea concentration are required for the score whereas surely other variables such as poor urine output, metabolic acidosis and clotting abnormalities will have a significant impact on morbidity (post-operative infection, wound healing etc) and mortality?

Actual data collection is a further potential problem that has been identified with POSSUM scoring generally. The timing of scoring the physiological variables is not standardised but is important as they may change with time. Reviewing the literature, it appears that recorded data in some studies has been collected on the patient’s admission and represents the worst score, whereas other workers have used data obtained immediately before surgery, prior to which a period of aggressive resuscitation may have taken place and the numbers may have improved. Missing data points are a common problem in all these studies. This may cause discrepancies in the outcome.

Some of the physiological investigations required to complete the data set are subject to misinterpretation. ECGs may be given artificial scoring if the reader detects a non-specific change but records the result as ‘any other’. Other parameters such as degree of cardiac failure, blood loss and peritoneal soiling may be wrongly estimated. A sea change in the way these parameters are recorded is vital to obtain an accurate clinical picture, and inform surgical management. Without this, any scoring system is considerably diminished in value.

How can CR-POSSUM be used to improve patient outcome?
We wonder if a different surgical approach is required for the peritonitic high-risk patient with a high predicted mortality by (accurately recorded) CR-POSSUM.

In trauma surgery “unstable patients nearing physiological exhaustion require damage control surgical tactics.”²⁸ The physiological derangement seen in trauma patients is due to a post-inflammatory state secondary to immune activation superimposed on traumatic injury. This may be indistinguishable at a cellular level from the mechanism of organ dysfunction that develops due to peritonitis, therefore a similar surgical approach may be appropriate.²⁸

Trauma surgeons select high-risk patients on the basis of acidaemia (pH<7.2 or base deficit >8), hypothermia (temperature <34 C) or diffuse coagulopathy. These patients will undergo “abbreviated surgery” to reverse shock and restore physiological stability but extensive reconstructive surgery is delayed until homeostasis is restored. For abbreviated laparotomy, primary abdominal closure may be delayed. Case studies suggest that this is an effective management strategy provided that there is an acceptance of repeated surgery.^{29, 30, 31}

A similar approach may be beneficial in some similar patients with peritonitis, particularly those with a perforated viscus secondary to malignancy and a high predicted mortality. These patients may do better with a damage-control laparotomy rather than an extensive resection procedure when they are already at the edge of their physiological boundary.³² Early data from some American centres strongly suggest this approach is beneficial, but performing a proper large scale clinical trial requires the recruitment of multiple surgical centres. This in turn requires a rigorously applied, standardised scoring system for patients with an acute abdomen. CR-POSSUM has the potential to be such a scoring system, but our surgical colleagues will require our help and expertise with well-validated systems such as APACHE to make this possible.

References

1. Knaus WA, Zimmerman JE, Wagner DP, Draper EA, Lawrence DE. APACHE – acute physiology and chronic health evaluation: a physiologically based classification system. Crit Care Med 1981; 9: 591-7
2. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE 2: a severity of disease classification system. Crit Care Med 1985; 13: 514-20
3. Jones HJS, de Cossart L. Risk scoring in surgical patients. Br J Surg 1999; 86: 149-157
4. Wagner D, Draper E, Knaus WA. Development of APACHE 3. Crit Care Med 1989; 17: S199-203
5. Wong DT, Knaus WA1991. Can J Anesth; 38: 374-83
6. Knaus WA, Wagner DP, Draper EA, Zimmerman JE. Prognosis in acute organ-system failure. Ann Surg 1985; 202: 685-93
7. Goris RJA, Boekhorst TPA, Nuytinck JKS, Gimbrere JSF. Multi-organ failure. Generalised autodestructive inflammation? Arch Surg 1985; 125: 1109-15
8. Ruokonen E, Takala J, Kari A, Alhava E. Septic shock and multi-organ failure. Crit Care Med 1991; 19:1146-51
9. Kollef MH, Allen BT, Determinants of outcome for patients in the intensive care unit requiring abdominal surgery: a prospective, single centre study. Chest 1994; 23: 1822-8
10. Ohmann C, Dietmar H, Hannes W. Prospective evaluation of prognostic scoring systems in peritonitis. Eur J Surg 1993; 159: 267-274
11. Bosscha K, Reijnders K, Hulstaert P F, Algra A, van der Werken C. Prognostic scoring systems to predict outcome in peritonitis and intra-abdominal sepsis Br J Surg 1997; 84: 1532-1534
12. Gruau G, Heemken R, Hau T. predictors of outcome in patients with postoperative intra-abdominal infection Eur J Surg 1996; 162: 619-625
13. Sutton R, Bann S, Brooks M, Sarin S. The surgical risk scale as an improved tool for risk-adjusted analysis in comparative surgical audit. Br J Surg 2002; 89: 763-768
14. Pacelli F, Doglietto GB, Alfieri S, Piccione E, Sgadari A, Gui D, Crucitti F. Prognosis in intra-abdominal infections: multivariate analysis on 604 patients. Arch Surg 1996; 131: 641-645
15. Neary WD, Prytherch CF, Heather BP, Earnshaw JJ. Comparison of different methods of risk stratification in urgent and emergency surgery. Br J Surg 2007; 94: 1300-1305
16. Horiuchi A, Watanabe Y, Doi T, Sato K, Yukumi S, Yoshida M, Yamamoto Y, Sugishita H, Kawachi K. Evaluation of prognostic factors and scoring system in colonic perforation. World J Gastroenterol 2007; 13(23):3228-3231
17. Neary WD, Heather BP, Earnshaw JJ. The physiological and operative severity score for the enumeration of mortality and morbidity. Br J Surg 2003; 90: 157-165
18. Jones DR, Copeland GP, de Cossart L. Comparison of POSSUM with APACHE 2 for prediction of outcome from a surgical high-dependency unit. Br J Surg 1992; 79: 1293-1296
19. Whitely MS, Prytherch DR, Higgins B, Weaver PC, Prout WG. An evaluation of the POSSUM scoring system. Br J Surg 1996; 83: 812-815
20. Constantinides VA, Tekkis PP, Senapati A. Comparison of POSSUM scoring system and surgical risk scale in patients undergoing surgery for complicated diverticular disease. Dis Colon Rectum 2006; 49: 1322-1331
21. Mohil RS, Bhartnagar LB, Rajneesh DKD, Magan M. POSSUM and p-POSSUM for risk adjusted audit of patients undergoing emergency laparotomy. Br J Surg 2004; 91: 500-503
22. Brooks MJ, Sutton R, Sarin S. Comparison of surgical risk score, POSSUM and p-POSSUM in higher risk surgical patients. Br J Surg 2005; 92: 1288-1292
23. Prytherch DR, Whiteley MS, Higgins B, Weaver PC, Prout WG, Powell SJ. POSSUM and p-POSSUM for predicting mortality. Br J Surg 1998; 85: 1217-1220
24. Tekkis PP, Prytherch DR, Kocher HM, Senapati A, Poloniecki JD, Stamatakis JD, Windsor ACJ. Development of a dedicated risk-adjusted scoring system for colorectal surgery (colorectal POSSUM). Br J Surg 2004; 91: 1174-1182
25. Vather R, Zargar-Shoshtari K, Adegbola S, Hill AG. Comparison of the POSSUM, p-POSSUM and CR-POSSUM scoring systems as predictors of postoperative mortality in patients undergoing major colorectal surgery. ANZ J Surg 2006; 76: 812-816
26. Ramkumar T, Ng V, Fowler L, Farouk R. A comparison of POSSUM, p-POSSUM and CR-POSSUM for the prediction of postoperative mortality in patients undergoing colorectal resection. Dis Colon Rectum 2006; 49: 330-335
27. Markus PM, Martell J, Leister I, Horstmann O, Brinker J, Becker H. Predicting postoperative morbidity by clinical assessment. Br J Surg 2005; 92: 101-106
28. Kirkpatrick AW, Ball CG, D’Amours SK, Zygun D. Acute resuscitation of the unstable trauma patient: bedside diagnosis and therapy. Can J Surg 2008; 51: 57-69
29. Fabian TC. Damage control in trauma: laparotomy wound management acute to chronic. Surg Clin N Am 2007; 87: 73-93
30. Schecter WP, Ivatury RR, Rotondo MF, Hirshberg A. Open abdomen after trauma and abdominal sepsis: a strategy for management. J Am Coll Surg 2006; 203(3) 390-396
31. Fantus RJ, Mellett MM, Kirby JP. Use of controlled fascial tension and an adhesion preventing barrier to achieve delayed primary fascial closure in patients managed with an open abdomen. Am J Surg 2006; 192: 243-247
32. Kulah B, Ozmen M, Ozer MV, Oruc MT, Coskun F. Outcomes of emergency surgical treatment in malignant bowel obstructions. Hepatogastroenterology 2005; 52: 1122-27