Risk-tailored prophylaxis for postoperative nausea and vomiting: has the big little problem gotten any smaller?
2017; Elsevier BV; Volume: 120; Issue: 1 Linguagem: Inglês
10.1016/j.bja.2017.11.005
ISSN1471-6771
Autores Tópico(s)Cardiac, Anesthesia and Surgical Outcomes
ResumoOne would think we are very close to a final solution for the prevention of postoperative nausea and vomiting (PONV). Our prophylactic antiemetics have been thoroughly studied in large trials and meta-analyses,1Apfel C.C. Korttila K. Abdalla M. et al.A factorial trial of six interventions for the prevention of postoperative nausea and vomiting.N Engl J Med. 2004; 350: 2441-2451Crossref PubMed Scopus (1144) Google Scholar, 2Carlisle J. Stevenson C.A. WITHDRAWN: Drugs for preventing postoperative nausea and vomiting.Cochrane Database Syst Rev. 2017; 7 (CD004125)Crossref PubMed Scopus (67) Google Scholar we have developed PONV prediction models to identify which patients need antiemetics,3van den Bosch J.E. Moons K.G. Bonsel G.J. Kalkman C.J. Does measurement of preoperative anxiety have added value for predicting postoperative nausea and vomiting?.Anesth Analg. 2005; 100: 1525-1532Crossref PubMed Scopus (102) Google Scholar, 4van den Bosch J.E. Kalkman C.J. Vergouwe Y. et al.Assessing the applicability of scoring systems for predicting postoperative nausea and vomiting.Anaesthesia. 2005; 60: 323-331Crossref PubMed Scopus (73) Google Scholar, 5Apfel C.C. Läärä E. Koivuranta M. Greim C.A. Roewer N. A simplified risk score for predicting postoperative nausea and vomiting: conclusions from cross-validations between two centers.Anesthesiology. 1999; 91: 693-700Crossref PubMed Scopus (1374) Google Scholar, 6Koivuranta M. Laara E. A survey of postoperative nausea and vomiting.Anaesthesia. 1998; 53: 413-414PubMed Google Scholar and we have tested several decision support tools that remind and motivate us to apply such risk-tailored strategies in clinical practice.7Kooij F.O. Klok T. Hollmann M.W. Kal J.E. Automated reminders increase adherence to guidelines for administration of prophylaxis for postoperative nausea and vomiting.Eur J Anaesthesiol. 2010; 27: 187-191Crossref PubMed Scopus (53) Google Scholar, 8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar, 9Kappen T.H. Moons K.G.M. van Wolfswinkel L. Kalkman C.J. Vergouwe Y. van Klei W.A. Impact of risk assessments on prophylactic antiemetic prescription and the incidence of postoperative nausea and vomiting: a cluster-randomized trial.Anesthesiology. 2014; 120: 343-354Crossref PubMed Scopus (39) Google Scholar, 10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar (The Cochrane Meta-analysis by Carlisle et al.2Carlisle J. Stevenson C.A. WITHDRAWN: Drugs for preventing postoperative nausea and vomiting.Cochrane Database Syst Rev. 2017; 7 (CD004125)Crossref PubMed Scopus (67) Google Scholar was withdrawn July 2017, because of the withdrawal of several articles that were included in the review. Further study is being done as a consequences of these withdrawals. For the purpose of this editorial, the withdrawal does not invalidate the conclusion.) We just need to decide which of those strategies works best. According to a report by DeWinter and colleagues11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar in this issue, we are not there yet. Unsatisfied with the results of complex risk-tailored strategies, they implemented a simplified algorithm to advise on PONV prophylaxis with great success. They documented the largest decrease in PONV incidence (10%) for any PONV impact study, with a large corresponding increase in administered prophylactic antiemetics (on average more than one additional antiemetic per patient). Let us review the study and compare its methods and results to other PONV impact studies before we conclude whether or not we should all adapt this simplified strategy. The authors implemented a departmental guideline for PONV prophylaxis that simplifies PONV risk estimation. They already had a guideline in place that used a formal prediction model based on several risk factors to decide on the number of prophylactic interventions. A low compliance to that guideline triggered them to simplify the algorithm by making the decision solely dependent on being a woman (three interventions) or a man (two interventions). The results were spectacular. The anaesthesiologists administered on average more than one additional antiemetic during the intervention audit than they did during the control audit. That is a much larger increase than most previous studies on PONV guideline implementation. However, the previous studies implemented a formal risk model as an automated decision support system, whereas the present study only provided the algorithm to the anaesthesiologists. The large increase in prophylaxis may be the result of the ease of use of a non-automated algorithm: when the algorithm is not automated, it should be simple. Nonetheless, for many hospitals, automated decision support is not yet available. In such situations, simplifying the PONV prophylaxis algorithm may be the perfect compromise between a risk-tailored strategy and a strategy where all patients receive the same prophylaxis. The audit study design that the authors used is quick and cost-efficient. The authors sampled antiemetic prescription and PONV occurrence over two 5-day periods before and after the intervention implementation, and were rewarded with plenty of insight into the successes of their implementation strategy. That is more efficient than a previous 12 000-patient cluster-randomized trial with a limited impact on antiemetic prescription and PONV.9Kappen T.H. Moons K.G.M. van Wolfswinkel L. Kalkman C.J. Vergouwe Y. van Klei W.A. Impact of risk assessments on prophylactic antiemetic prescription and the incidence of postoperative nausea and vomiting: a cluster-randomized trial.Anesthesiology. 2014; 120: 343-354Crossref PubMed Scopus (39) Google Scholar In contrast, non-randomized studies always run the risk of unmeasured residual confounding due to variations in time. For example, surgical scheduling is infamous for its variability.12Pariser J.J. Diamond A.J. Christianson L.W. Mitchell B.A. Langerman A. Operating room inefficiencies attributable to errors in surgical case scheduling and surgeon procedure heterogeneity.Am J Med Qual. 2016; 31: 584-588Crossref PubMed Scopus (6) Google Scholar As only 5 consecutive working days were sampled, it is not unreasonable to expect large differences in PONV rates over time. In addition, provider scheduling also has its variabilities: 25% of the attending anaesthesiologists and 70% of the residents were not working during both audit weeks. Even though the increased antiemetic prescription was still most likely caused by the intervention, it is conceivable that a group of enthusiasts–only working during the intervention audit–caused the increase in antiemetic prescription. This illustrates an important difference with a classic drug-based intervention study: guidelines, algorithms, and decision support often aim to inform healthcare providers without directly affecting patients. It is therefore prudent to have insight into the baseline characteristics of the healthcare providers in each of the study groups. Which PONV prophylaxis strategy of the different impact studies would be the best strategy? This requires a comparison of PONV incidences, antiemetic administration, and how risk-tailored the strategies are. For a proper comparison of the four main impact studies and their strategies,8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar, 9Kappen T.H. Moons K.G.M. van Wolfswinkel L. Kalkman C.J. Vergouwe Y. van Klei W.A. Impact of risk assessments on prophylactic antiemetic prescription and the incidence of postoperative nausea and vomiting: a cluster-randomized trial.Anesthesiology. 2014; 120: 343-354Crossref PubMed Scopus (39) Google Scholar, 10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar, 11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar Table 1 presents their results extrapolated to the global yearly volume of surgical procedures (313 million procedures).13Weiser T.G. Haynes A.B. Molina G. et al.Estimate of the global volume of surgery in 2012: an assessment supporting improved health outcomes.Lancet. 2015; (S11)Google ScholarTable 1PONV prophylaxis strategies extrapolated to the global yearly volume of surgical procedures (313 million procedures). The overall results are based on the reported data. For risk-subgroups numbers were calculated using the average number of patients in that subgroup across both intervention and control groups. Hence, the numbers of both risk-subgroups might not exactly add up to the overall results.DeWinter and colleagues11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google ScholarKooij and colleagues8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar,∗Kooij and collegaues8 reported the administration of specific antiemetics, rather than the number of antiemetics. In this table, it was assumed that dexamethasone was always the first antiemetic of choice, making the percentage of patients that received granisetron the percentage of patients that received two antiemetics.Kappen and colleagues9Kappen T.H. Moons K.G.M. van Wolfswinkel L. Kalkman C.J. Vergouwe Y. van Klei W.A. Impact of risk assessments on prophylactic antiemetic prescription and the incidence of postoperative nausea and vomiting: a cluster-randomized trial.Anesthesiology. 2014; 120: 343-354Crossref PubMed Scopus (39) Google ScholarKappen and colleagues10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google ScholarLow-riskAll patientsHigh-riskLow-riskAll patientsHigh-riskLow-riskAll patientsHigh-riskLow-riskAll patientsHigh-risk≤2 RF≥3 RF≤2 RF≥3 RFrisk <60%risk ≥60%risk <60%risk ≥60%n(%)164×106 (52)313×106149×106 (48)216×106 (69)313×10697×106 (31)293×106 (94)313×10620×106 (6)293×106 (94)313×10620×106 (6)PONV PONV in control group (Y0), n (%)37×106 (23)102×106 (33)67×106 (45)42×106 (19)87×106 (28)46×106 (48)120×106 (41)134×106 (43)14×106 (70)141×106 (48)158×106 (50)16×106 (82) PONV in intervention group (Y1), n (%)29×106 (18)70×106 (22)40×106 (27)42×106 (19)72×106 (23)30×106 (31)114×106 (39)127×106 (41)13×106 (69)123×106 (42)132×106 (42)10×106 (50) Absolute PONV risk reduction, n (pp)8×106 (5)32×106 (10)26×106 (18)0×106 (0)15×106 (5)16×106 (17)6×106 (2)7×106 (2)0×106 (1)17×106 (6)25×106 (8)6×106 (32)Antiemetic prophylaxisAverage no. of prophylactic antiemetics per patient Control group (X0), mean0.661.071.550.410.751.580.180.180.300.280.300.68 Intervention group (X1), mean2.012.232.460.280.761.780.320.340.790.920.961.90Absolute increase in prophylactic antiemetics, n221×106363×106136×106−28×1063×10619×10641×10650×10610×106188×106207×10624×106Change in prophylaxis (intervention - control), n (pp) Patients receiving 0 prophylactic antiemetics−81×106 (−49)−111×106 (−35)−28×106 (−19)13×106 (6)−2×106 (−1)−9×106 (−10)−33×106 (−11)−40×106 (−13)−8×106 (−38)−123×106 (−42)−128×106 (−41)−6×106 (−31) Patients receiving 1 prophylactic antiemetic−28×106 (−17)−62×106 (−20)−33×106 (−22)1×106 (1)0×106 (0)−1×106 (−1)26×106 (9)31×106 (10)6×106 (29)58×106 (20)54×106 (17)−5×106 (−26) Patients receiving 2 prophylactic antiemetics87×106 (53)112×106 (36)25×106 (17)−14×106 (−7)2×106 (1)10×106 (11)7×106 (2)8×106 (3)2×106 (8)66×106 (22)69×106 (22)4×106 (22) Patients receiving ≥3 prophylactic antiemetics22×106 (13)61×106 (19)36×106 (24)n/an/an/a1×106 (0)1×106 (0)0×106 (1)0×106 (0)4×106 (1)7×106 (35)Additional antiemetics per prevented PONV case28115< 00.3177481184Severe adverse events (as a result of the strategy)Assumed incidence 1:100 000 per antiemetic Severe adverse events in control group, n1084333723108802344153151557459815952132 Severe adverse events in intervention group, n32976990366260323811728932107515526993013371 Increase in severe adverse events, n221336531351−277361974175019618842061239Assumed incidence 1:10 000 per antiemetic Severe adverse events in control group, n10 84133 36623 104879923 44315 30851535741591815095221320 Severe adverse events in intervention group, n32 97169 89736 618602923 80617 282931810 752155226 99430 1293708 Increase in severe adverse events, n22 13036 53013 515−276936319754165501196118 84420 6072388PONV, postoperative nausea and vomiting; pp, percentage point; RF, risk factors.∗ Kooij and collegaues8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar reported the administration of specific antiemetics, rather than the number of antiemetics. In this table, it was assumed that dexamethasone was always the first antiemetic of choice, making the percentage of patients that received granisetron the percentage of patients that received two antiemetics. Open table in a new tab PONV, postoperative nausea and vomiting; pp, percentage point; RF, risk factors. The present study reported a 10% decrease in PONV incidence and an average increased administration of more than one prophylactic antiemetic per patient. Worldwide, the present study would result in the largest decrease in PONV incidence of the four studies (32 million prevented PONV cases), with an additional 363 million antiemetics administered to patients. This is substantially more than the runner up, with 25 million prevented PONV cases and an additional 207 million prophylactic antiemetics administered.10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar However, a risk-tailored approach to PONV prophylaxis is not about sheer volume. A risk-tailored approach aims to be efficient: the largest decrease in PONV with the lowest administration of prophylactic antiemetics. In other words, administer more prophylaxis to patients who will benefit the most from prophylaxis (high-risk patients) and less to low-risk patients. For example, Kooij and colleagues8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar may have the lowest overall numbers in antiemetic administration (Table 1), yet they were very efficient in risk-tailored prophylaxis. In their high-risk group (three or more risk factors), there would be an increase of 19 million administered antiemetics, whereas in their low-risk group (from zero to two risk factors), their strategy would result in decreased administration of 28 million antiemetics. The present study by DeWinter and colleagues11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar has much less contrast between risk groups, which confirms that the simplified strategy leans towards a strategy where all patients receive the same prophylaxis. The two strategies in the studies of Kappen and colleagues9Kappen T.H. Moons K.G.M. van Wolfswinkel L. Kalkman C.J. Vergouwe Y. van Klei W.A. Impact of risk assessments on prophylactic antiemetic prescription and the incidence of postoperative nausea and vomiting: a cluster-randomized trial.Anesthesiology. 2014; 120: 343-354Crossref PubMed Scopus (39) Google Scholar, 10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar would result in an increase in antiemetic prophylaxis in both low-risk and high-risk patients. There is a difference in the PONV risk profiles between all studies. The studies of DeWinter and colleagues11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar and Kooij and colleagues8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar used the Apfel score to predict PONV risk.5Apfel C.C. Läärä E. Koivuranta M. Greim C.A. Roewer N. A simplified risk score for predicting postoperative nausea and vomiting: conclusions from cross-validations between two centers.Anesthesiology. 1999; 91: 693-700Crossref PubMed Scopus (1374) Google Scholar If we follow Kooij and colleagues8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar, we would classify three or more risk factors as high risk, which would be equivalent to a predicted risk of 60% or greater. In these two studies, approximately a one-third to one-half of patients is high risk according to that definition. The two studies by Kappen and colleagues9Kappen T.H. Moons K.G.M. van Wolfswinkel L. Kalkman C.J. Vergouwe Y. van Klei W.A. Impact of risk assessments on prophylactic antiemetic prescription and the incidence of postoperative nausea and vomiting: a cluster-randomized trial.Anesthesiology. 2014; 120: 343-354Crossref PubMed Scopus (39) Google Scholar, 10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar used a modified Van den Bosch model by which only 6% of patients had a high predicted PONV risk.3van den Bosch J.E. Moons K.G. Bonsel G.J. Kalkman C.J. Does measurement of preoperative anxiety have added value for predicting postoperative nausea and vomiting?.Anesth Analg. 2005; 100: 1525-1532Crossref PubMed Scopus (102) Google Scholar, 14Kappen T.H. Vergouwe Y. van Klei W.A. van Wolfswinkel L. Kalkman C.J. Moons K.G.M. Adaptation of clinical prediction models for application in local settings.Med Decis Making. 2012; 32 (E1–10)Crossref PubMed Scopus (38) Google Scholar The observed PONV incidences in high-risk patients in the studies of Kappen and colleagues9Kappen T.H. Moons K.G.M. van Wolfswinkel L. Kalkman C.J. Vergouwe Y. van Klei W.A. Impact of risk assessments on prophylactic antiemetic prescription and the incidence of postoperative nausea and vomiting: a cluster-randomized trial.Anesthesiology. 2014; 120: 343-354Crossref PubMed Scopus (39) Google Scholar, 10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar appears to be higher (70–82% PONV in high-risk patients in the control group) than in the other two studies (45–48% PONV). Although there may be differences in study populations, a very plausible explanation for the differences in risk distribution is the use of different prediction models. Despite a large increase in the average number of prophylactic antiemetics per high-risk patient, the strategy of Kappen and colleagues10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar was more restrictive in antiemetic administration in high-risk patients than the strategies by DeWinter and colleagues11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar and Kooij and colleagues,8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar simply because the prediction model classified fewer patients into the high-risk group. Now that we have a concept of how different strategies would treat patients with varying risks, we need to evaluate how patients in different risk groups would benefit from these strategies. In high-risk patients, the strategy by DeWinter and colleagues11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar would prevent 26 million cases of PONV, followed by Kooij and colleagues8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar with 16 million and Kappen and colleagues10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar with six million prevented cases. When we weigh the 'costs' (antiemetic administration) against the 'benefits' (PONV prevention), that ranking changes. The strategy by Kooij and colleagues8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar would require only one additional antiemetic per prevented PONV case (19 million additional antiemetics to prevent 16 million PONV cases), followed by Kappen and colleagues10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar with four additional antiemetics per case (24 million antiemetics to prevent 6 million PONV cases), and DeWinter and colleagues11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar would need five additional antiemetics per prevented PONV case (136 million antiemetics to prevent 26 million PONV cases). It should be noted that the impact on PONV prevention may be somewhat overestimated in the strategy by Kooij and colleagues8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar: if one additional antiemetic would always prevent one case of PONV in high-risk patients, the incidence of PONV in high-risk patients should be zero rather than 31%. This is probably a side effect of a before-after study design. Note that this does not mean that the studies of DeWinter and colleagues11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar and Kappen and colleagues10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar did not suffer from such overestimation, it is only less apparent. In low-risk patients, the strategies by DeWinter and colleagues11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar and Kappen and colleagues10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar are much less efficient. DeWinter and colleagues11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar would result in 221 million additional antiemetics to prevent 8 million cases of PONV (28 additional antiemetics per prevented case), whereas Kappen and colleagues10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar would administer 188 million additional antiemetics for the benefit of 17 million patients (11 additional antiemetics per prevented case). The strategy of Kooij and colleagues8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar is even more efficient in low-risk patients: a reduction of 28 million antiemetics without an increase in PONV, resulting in less than zero additional antiemetics per prevented case. Most likely, this is also a side effect of the before-after study design. Despite the possibility of an overestimated effect, the strategy by Kooij and colleagues8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar is likely to yield the most efficient change in PONV prophylaxis administration. The studies only provide an answer to which strategy would result in the most efficient change in antiemetic prophylaxis, not which strategy would overall be the most efficient. After all, all studies did not implement their strategy in an 'untouched population': even before the implemented strategy, patients already received risk-based prophylaxis. We would like to compare strategies for their total numbers of prophylactic antiemetics that would be administered in relation to their total number of patients who suffered from PONV. Unfortunately, the differences in the risk profiles between patient populations prohibit such comparison. Summarizing, the more liberal approach to antiemetic prophylaxis by DeWinter and colleagues11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar might lead to the largest decrease in PONV, but it is not a very efficient strategy with an average of 11 additional antiemetics required to prevent a single PONV case. Whether the PONV decrease justifies the additional antiemetics depends on the negative impact of the additional antiemetics. Financial costs aside, we know that commonly used antiemetics can be considered safe in thousands of patients. However, in this editorial we extrapolated the strategy to hundreds of millions of patients. Suppose the incidence of severe adverse events per prophylactic antiemetic would be one in 100 000, we would expect that the strategy by DeWinter and colleagues11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar would cause an additional 3653 patients to suffer from a severe adverse event. Note that these numbers should only be considered a thought experiment and not a true result. If the incidence per antiemetic were one in 10 000, an additional 36 530 patients would suffer from a severe adverse event, resulting in a total of almost 70 000 patients. As a percentage, that is low, but in absolute numbers it might not be. An individual patient might consider a risk of one in 10 000 very acceptable. However, is it acceptable to us as professionals to potentially harm thousands of patients to prevent a short-lasting event such as PONV in millions of patients? The big little problem has just gotten a little bigger.15Kapur P.A. The big 'little problem'.Anesth Analg. 1991; 73: 243-245Crossref PubMed Google Scholar After all the research that has been done on PONV prophylaxis, it appears that we are still short on information. The ideal solution would be to perform some very large multicentre studies to find the adverse event rates for our antiemetics. As that would require so many patients and a more complicated follow-up survey than simply asking for PONV, it would be very expensive. Even though that would bring us much closer to a final answer, it may be unjustifiable to spend that much effort and public money on our big little problem. Despite the success of the strategy by DeWinter and colleagues,11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar a risk-tailored strategy for PONV would do the most justice to our patients, as long as we do not know the true incidence of adverse events due to antiemetics. However, there are still several opportunities to improve such risk-tailored strategies, even without performing that large multicentre trial. First, we could further optimize our risk-dependent algorithms. The current impact studies suggest that increased administration of antiemetics in very high-risk patients (Kappen and colleagues10Kappen T.H. Vergouwe Y. van Wolfswinkel L. Kalkman C.J.J. Moons K.G.M.G.M. van Klei W.A.A. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting.Br J Anaesth. 2015; 114: 252-260Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar) and a more restrictive administration in low-risk patients (Kooij and colleagues8Kooij F.O. Vos N. Siebenga P. Klok T. Hollmann M.W. Kal J.E. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population.Br J Anaesth. 2012; 108: 961-965Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar) might result in a more optimal algorithm. Further study of various risk-dependent algorithms may help us to improve our Society for Ambulatory Anesthesia (SAMBA) guidelines.16Gan T.J. Diemunsch P. Habib A.S. et al.Consensus guidelines for the management of postoperative nausea and vomiting.Anesth Analg. 2014; 118: 85-113Crossref PubMed Scopus (920) Google Scholar Second, as not all patients respond to prophylactic antiemetics, it might be a worthwhile ambition to predict which patients will be a prophylaxis responder. It is easier to justify the risk of adverse events of prophylactic antiemetics when patients have a higher probability to respond to the antiemetics–especially if we can predict to which antiemetic they are most likely to respond. Third, even when we know which patient should receive how many and which prophylactic antiemetics, there will always be the problem of implementing such algorithms. PONV is important, but there are other pressing matters during anaesthesia cases.17Kappen T.H. van Loon K. Kappen M.A.M. et al.Barriers and facilitators perceived by physicians when using prediction models in practice.J Clin Epidemiol. 2016; 70: 136-145Abstract Full Text Full Text PDF PubMed Scopus (32) Google Scholar Implementation of decision support is not a PONV-specific problem, but we might be able to add to the understanding of how we can implement decision support to help ourselves to do the right thing. Awaiting further scientific answers, DeWinter and colleagues11DeWinter G. Staelens W. Veef E. Teunkens A. Van de Velde M. Rex S. A simplified algorithm for the prevention of postoperative nausea and vomiting: a before-and-after study.Br J Anaesth. 2018; 120: 156-163Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar have shown a feasible way to successfully implement a risk-tailored algorithm for PONV prophylaxis, especially when an infrastructure that enables more complex, automated decision support is not yet available. Manuscript concept and design; acquisition of data; statistical analysis; interpretation of data; drafting of the manuscript; critical revision of the manuscript for important intellectual content: T.H.K. None declared. This work had no financial support other than departmental funding.
Referência(s)