This Correspondence accompanies the following article:
Kranke P, Eberhart LH. Simple risk scoring for postoperative nausea and vomiting may work, but that does not necessarily make patients less sick! Eur J Anaesthesiol 2012; 29:59–60.
We would like to contribute our experience in implementing prophylactic strategies to the debate on postoperative nausea and vomiting (PONV) that was recently published. This is a highly complex area, reflected by the continuing difficulty in translating evidence into improved patient care as outlined by Kranke.1 Our experience over 5 years of risk stratification and targeted prophylaxis shows some success but also raises questions about how to benefit the most and harm the least. Our data from single centre clinical audits performed on a near annual basis between 2006 and 2010 cast light upon the trials, tribulations and practicalities of trying to reduce PONV in daily practice.
Eberhart and Morin2 dispute the use of risk stratification for predicting PONV because of its inability to make a valid individual assessment. But this is a misunderstanding of a risk scoring process that estimates the probability of a discrete outcome. By definition, such scores estimate risk and cannot possibly predict individual outcome. Their use, therefore, is to stratify patients so that those at highest risk may be identified and targeted for the most aggressive management. In that respect, we can confirm that one simplified scoring system does work in routine practice.
Our centre has operated risk adaptive guidelines for PONV prophylaxis since 2006. Risk is based upon the Koivuranta score (K), selected because of its simplicity, objectivity and superior calibration compared with Apfel.3 We have simplified this further into just three groups for purposes of prophylaxis: low risk (K = 0–1), medium risk (K = 2–3) and high risk (K = 4–5). Our guidelines recommend that high-risk patients receive at least three antiemetics, medium risk at least two and low risk zero or one.
Cumulative results of our audits are shown in the table (Table 1) and confirm that simplified scoring effectively categorises patients by risk of PONV, producing groups with clinically and statistically significant differences in PONV (P < 0.05, by analysis of confidence intervals). They also confirm that the high-risk and medium-risk groups benefit from our targeted guideline.
Unfortunately, however, there is not yet a demonstrable overall effect upon PONV because adherence to the guidelines has been persistently low. Only 37% of medium-risk and high-risk patients receive the specified prophylaxis, a finding which certainly supports some of the arguments that Eberhart and Morin2 were making. However, we share Pierre's stance4 against abandoning attempts at risk stratification and are against the universal use of two-agent PONV prophylaxis as promoted by Eberhart and Morin. First, giving two antiemetics to all patients in ignorance of their underlying PONV risk would deny high-risk patients the potential benefit from a third or fourth agent. Second, many low-risk patients would be exposed to possible side-effects of multiple agents without benefit. As an example, if the findings of Apfel et al.5 are extrapolated to patients with low risk of PONV, a second antiemetic would reduce PONV risk in our low-risk patients from 16 to 12%. This gives a number needed to treat of 25, which means 24 patients would be treated without benefit but exposed to risks of side-effects. Contrary to Eberhart and Morin's opinion, we do not believe that these side-effects are all benign and nor would patients, such as those suffering ‘locked in’ sensations after droperidol. We suggest that in low-risk patients, prophylaxis should be restricted and PONV treated if it occurs.
Our audit data demonstrate that a simple three-tier risk scoring system can classify patients by risk of PONV, that medium-risk to high-risk patients can then benefit from multimodal prophylaxis and low-risk patients need not receive drugs from which there is little likelihood of benefit. PONV is never going to be eliminated by current management options. We should concentrate our efforts, and audit, upon optimising patients’ management based upon risk and minimising the incidence and severity of PONV with both prophylaxis and treatment.
None of the authors has any conflict of interest.
1. Kranke P. Effective management of postoperative nausea and vomiting: let us practice what we preach!. Eur J Anaesthesiol
2. Eberhart LHJ, Morin AM. Risk sores for predicting postoperative nausea and vomiting are clinically useful tools and should be used in every patient: con – life is really simple, but we insist upon making it complicated. Eur J Anaesthesiol
3. Van den Bosch JE, Kalkman CJ, Vergouwe Y, et al. Assessing the applicability of scoring systems for predicting postoperative nausea and vomiting. Anaesthesia
4. Pierre S. Risk scores for predicting postoperative nausea and vomiting are clinically useful tools and should be used in every patient: pro – ‘don’t throw the baby out with the water’. Eur J Anaesthesiol
5. Apfel CC, Kortilla K, Abdalla M, et al. A factorial trial of six interventions for the prevention of postoperative nausea and vomiting. N Engl J Med