The Relationship Between Patient Risk Factors and EarlyVersus Late Postoperative Emetic Symptoms

Paul F. White, PhD, MD,FANZCA

Ozlem Sacan, MD

Nina Nuangchamnong, MS

Tiffany Sun, MS

Matthew R. Eng, MS

BACKGROUND: Postoperative nausea and vomiting (PONV) remain common compli-cations after elective surgery. Prophylactic antiemetic drugs are frequently admin-istered to patients with well known risk factors for developing PONV. Wedesigned this prospective observational study to assess the relationship betweencommon patient risk factors for developing PONV and the occurrence of early(0–24 h) versus late (24–72 h) emetic symptoms.METHODS: One hundred thirty patients undergoing elective laparoscopic (n � 88) orplastic (n � 42) surgery were assigned a risk score for developing PONV based onthe Apfel risk scoring system, which assigns one point each for female gender,nonsmoking status, history of PONV or motion sickness, and postoperative opioiduse. It was assumed that all patients would receive an opioid analgesic in thepostoperative period. The patients received 0, 1, 2, or 3 antiemetic drugs forprophylaxis. The occurrence of nausea, vomiting, and need for rescue antiemeticswas assessed at specific time intervals from 0 to 6, 6–24, and 24–72 h after surgery.In addition, the impact of PONV on recovery of normal activities of daily livingwas assessed using a standardized patient questionnaire.RESULTS: One or more prophylactic antiemetics were administered to 87%, 90%, and95% of the patients in the two, three, and four Apfel risk-factor groups, respectively. Inthe presence of three or four risk factors, �2 antiemetics were administered to 56% and75% of the patients, respectively. Vomiting was reported in 11% and 22% of patients inthe three and four risk factor groups compared with 6% in the two risk factor group at0–6 h, and 13% and 27% (vs 0%) at 6–24 h, respectively. However, in the 24–72 hpostoperative period, the incidences of emesis were low and did not differ in the threerisk groups (9%, 5%, and 11%, respectively). The occurrence of moderate-to-severenausea was increased in the higher risk groups at 0–6 h and 6–24 h (19%–28% vs 6%and 20%–30% vs 9%, respectively). However, the incidences of nausea in the 24–72 hperiod in the three and four risk factor groups were not different from the two-riskfactor group (5% and 8% vs 6%, respectively). The need for rescue antiemetics andinterference of emetic symptoms with normal activities was greater in the four riskfactor group compared with the two and three risk factor groups.CONCLUSION: Despite the frequent use of multiple antiemetic drugs for prophylaxis, anApfel risk score of three or four (vs 2) was associated with a higher incidence of emeticsequelae in the first 24 h after surgery. However, the occurrence of late (24–72 h) emeticsymptoms was low and appeared to be unrelated to the patient’s Apfel risk score.(Anesth Analg 2008;107:459–63)

Postoperative nausea and vomiting (PONV) is along-standing problem for anesthesia practitioners.1

The incidence of PONV remains high despite thefrequent use of prophylactic antiemetics (e.g., 5-HT3

antagonists, glucocorticoids, dopamine antagonists),shorter-acting anesthetics and analgesics (e.g., propo-fol, desflurane, remifentanil), and less invasive surgi-cal techniques (e.g., laparoscopic procedures). Patient,anesthetic and surgical factors all contribute to thepersistently frequent incidence of emetic symptoms inthe postoperative period.1 With the increasingly em-phasis on earlier mobilization and discharge (“fast-tracking”) after both minor and major operations,2

postural hypotension and oral opioid containing an-algesics may become more important contributors toPONV. In a recent analysis of factors influencingpostanesthesia recovery, Edler et al.3 reported that thenumber of episodes of PONV contributes significantlyto prolonging the patient’s length of stay in thehospital.

From the Department of Anesthesiology and Pain Management,University of Texas Southwestern Medical Center at Dallas, Texas.

Accepted for publication April 4, 2008.Dr. Paul F. White, Section Editor for Special Project, was recused

from all editorial decisions related to this manuscript.Supported by endowment funds from the Margaret Milam

McDermott Distinguished Chair in Anesthesiology, and UT South-western Medical Student Summer Research Program, and the WhiteMountain Institute, a non-for profit private foundation (Dr. P.F.W.,President).

Address correspondence and reprint requests to Dr. Paul F.White, PhD, MD, FANZCA, Department of Anesthesiology andPain Management, University of Texas Southwestern Medical Cen-ter at Dallas, 5323 Harry Hines Boulevard, Dallas, Texas 75390-9068.Address e-mail to paul.white@utsouthwestern.edu.

Copyright © 2008 International Anesthesia Research SocietyDOI: 10.1213/ane.0b013e31817aa6e4

Vol. 107, No. 2, August 2008 459

Use of antiemetic prophylaxis has been shown toimprove patient satisfaction and speed of recoverycompared to simply treating the symptoms when theyoccur in the postoperative period.4–6 Therefore, anti-emetic drugs are now commonly administered both atthe start and end of surgery to patients considered tobe at increased risk of developing PONV.7 In fact,combinations of antiemetic drugs are now routinelyadministered as part of a multimodal strategy forreducing postoperative emetic symptoms in “at risk”patient populations.8–10 Apfel et al.11 have developeda simplified scoring system which has favorable dis-criminating and calibrating properties for predictingan individual patient’s risk for developing PONV.12

This prospective observational study was designedto assess the relationship between the patients’ Apfelrisk score for developing PONV and the occurrence ofearly (�24 h) and late (24–72 h) emetic symptoms in“at risk” surgical populations receiving frequent mul-timodal antiemetic prophylaxis.

METHODSOne hundred thirty healthy patients (18–75 yr)

undergoing major laparoscopic (n � 88) or plastic (n �42) surgery under general anesthesia were enrolled inthis IRB-approved prospective observational studyafter obtaining written, informed consent. Patientswere excluded from this study if they had receivedany antiemetic medications within 24 h of their sur-gery, had any unstable preexisting medical conditions,received an investigational medication within 30 oftheir surgery, or had a recent (�6 mo) history ofalcohol or drug abuse. All patients entered into thisstudy were expected to receive postoperative opioidanalgesics, and had at least one additional risk factor(i.e., minimum Apfel score of two). Patients with anApfel score of three or four were considered to be at“high risk” for developing PONV.

All patients received a standardized general anes-thetic technique consisting of IV midazolam 20 �g/kg,for premedication; propofol 1–2 mg/kg, fentanyl 1–2�g/kg, and rocuronium 0.6–0.8 mg/kg IV for induc-tion, and desflurane 4%–8% inspired in a 50:50air:oxygen mixture for maintenance of anesthesia.Residual neuromuscular blockade was reversed at theend of surgery with a combination of neostigmine 2–4mg IV, and glycopyrolate, 0.4–0.6 mg IV. Antiemeticprophylaxis was administered at the discretion of theattending anesthesiologist, and when administered,consisted of one or more of the following IV medica-tions: ondansetron 4 mg, dexamethasone 4 mg,and/or droperidol 0.625 mg. In the postanesthesiacare unit (PACU), pain relief was initially providedwith fentanyl, 25–50 �g IV. In the postdischargeperiod, patients were prescribed oral opioid-containing analgesics (e.g., Lortab or Vicodin) to treatmoderate-to-severe pain. Ondansetron, 4 mg IV, was

administered as the initial rescue antiemetic medica-tion in the PACU (unless it had been used for anti-emetic prophylaxis). Promethazine, 12.5 mg IV, wasused as the rescue antiemetic if ondansetron failed toprevent (or relieve) early postoperative emetic symp-toms. Prochlorperazine suppositories were prescribedfor treating PONV in the postdischarge period.

An emetic episode was defined as any number ofvomits (or retches) in a single 5 min period or asequence of vomits in close succession not relieved bya 2 min period of relaxation. If the patient complainedof nausea at any of the assessment intervals, they wereasked to quantify the severity of the nausea using afour-point Likert scale, with 0 � none, 1 � mild, 2 �moderate, 3 � severe. If the patient complained ofmoderate-to-severe nausea lasting �15 min or experi-enced repeated episodes of emesis (or retching), theywere given the option of receiving a “rescue” antiemetic.

Rescue antiemetic drugs were administered by thePACU and short-stay unit nurses who were unawareof the patients’ Apfel risk scores. Follow-up evalua-tions were performed by one of the investigators at 24,48, and 72 h intervals after surgery to inquire about theoccurrence of emetic symptoms and need for rescueanalgesic and/or antiemetic medications. Using astructured questionnaire format, all patients wereasked if the occurrence of PONV had any impact ontheir ability to resume normal activities of daily living(e.g., ability to consume liquids and solid food, ambu-late without assistance). Simple “yes or no” responsesto the questions were recorded). At the time of dis-charge, patients were instructed to contact the investi-gator directly if they experienced any emetic symptomsafter discharge from the hospital and were prescribedprochlorperazine suppositories if their emetic symp-toms persisted.

In evaluating these outcome data, patients wereassigned to one of three risk groups based on theirsimplified Apfel score.11 This validated risk scoringsystem assigns one point for each of the followingfactors: 1) female gender, 2) nonsmoking status, 3)history of PONV or motion sickness, and 4) use ofopioid analgesics in the postoperative period. Data arepresented as means (�sd) for continuous data orrelative and absolute frequencies for discrete (cat-egorical) data. For statistical analysis, Fisher’s exacttest was used to analyze all categorical data. A signifi-cant difference was defined as a P value �0.05 whencomparing data from the three different risk factorgroups.

RESULTSPatient demographic characteristics are summa-

rized in Table 1. Demographic characteristics for pa-tients with two risk factors were significantly differentfrom those in the Apfel three or four risk factor groups(e.g., larger proportion of women, nonsmokers andnondrinkers, history of PONV, and motion sickness).

460 PONV Risk Factors and Early vs Late Emesis ANESTHESIA & ANALGESIA

All patients received opioid analgesics in the postop-erative period. Prophylactic antiemetics were admin-istered to 88%, 90% and 95% of the patients in the two,three, and four risk factor groups, respectively (Table1). Of importance, �2 antiemetic drugs were admin-istered to 29%, 56% and 75% of the patients with two,three, and four risk factors, respectively. However,44% and 24% of the patients in the three and four riskgroups, respectively, received one (or zero) prophy-lactic antiemetic drugs (Table 2).

Postoperative vomiting was reported in 11% and22% in the three and four risk factor groups, respec-tively, compared to only 6% in the two risk factorgroup during the early (0–6 h) postoperative period

(Table 2). The occurrence of moderate or severe nau-sea at 0–6 h was also more frequent in the two higherrisk groups (28 and 19% vs 6%, respectively). Require-ments for rescue antiemetic in the 0–6 h postoperativeperiod were 13%, 20%, and 22% in the two, three, andfour risk factor groups, respectively. All the patients inthe three and four risk factor groups who requiredrescue antiemetics received either no or only oneprophylactic antiemetic drug.

In the first 24 h after surgery, symptoms of vomit-ing (6%, 24%, and 49%), moderate-to-severe nausea(15%, 48%, and 49%), and the need for rescue anti-emetics (19%, 35%, and 54%) increased in proportionto the number of Apfel risk factors (i.e., two, three, and

Table 1. Demographic and Baseline Characteristics of Patients with 2, 3 or 4 Risk Factors for Developing Postoperative Nauseaand Vomiting (PONV) According to the Apfel Scoring System11,12a

Risk factors for PONVb

Two (n � 32) Three (n � 61) Four (n � 37)Mean age, range (yr) 51 (32–68) 48 (32–64) 54 (41–67)Sex (M/F) (n) 4/28 16/45c 1/36c

No smoking history (%) 50 60 73c

No alcohol consumption (%) 19 54c 43c

History of PONV (%) 3 20c 68c

History of motion sickness (%) 3 21c 62c

Mean anesthesia time (min) 165 � 79 177 � 92 146 � 86Mean surgery time (min) 148 � 31 157 � 87 130 � 71M � male; F � female.a Values are means (range or �SD), numbers (n), or percentages (%).b Apfel scoring system (4 points): 1) female gender, 2) history of PONV or motion sickness, 3) non-smoking status, 4) use of postoperative opioid analgesics.c P � 0.05 vs 2 risk factor group.

Table 2. Use of Prophylactic Antiemetics and Incidences of Postoperative Vomiting or Retching and Moderate-to-Severe Nausea, aswell as Requirement for Rescue Antiemetic Drugs And Interference of Postoperative Nausea and Vomiting (PONV) in Return toNormal Activities in The Three Risk Groupsa

Risk factors for PONV (n)c

Two (n � 32) Three (n � 61) Four (n � 37)Prophylactic antiemetics (n, %)

Zero 4, 13 6, 10 2, 5b

One 19, 59 21, 34 7, 19Two 5, 16 15, 25 6, 16Three 4, 13 19, 31 22, 59b

Vomiting or retching (n, %)0–6 h interval 2, 6 7, 11 8, 22b

6–24 h interval 0, 0 8, 13 10, 27b

24–72 h interval 3, 9 3, 5 4, 11Moderate-severe nausea (n, %)c

0–6 h interval 2, 6 17, 28b 7, 196–24 h interval 3, 9 12, 20 11, 30b

24–72 h interval 2, 6 3, 5 3, 8Rescue antiemetic administered (n, %)

0–6 h interval 4, 13 12, 20 8, 22b

6–24 h interval 2, 6 9, 15 12, 32b

24–72 h interval 3, 9 5, 8 7, 19Emetic symptoms interfered with normal activities (n, %)

0–6 h interval 1, 3 5, 8 6, 16b

6–24 h interval 0, 0 8, 13 12, 32b

24–72 h interval 3, 9 3, 5 4, 11a Values are numbers (n), percentages (%).b P � 0.05 vs 2 risk factors.c Nausea severity scale: 0 � none, 1 � mild, 2 � moderate, 3 � severe.

Vol. 107, No. 2, August 2008 © 2008 International Anesthesia Research Society 461

four, respectively). However, in the late (24–72 h)postoperative period, there was no apparent relation-ship between PONV symptoms and the number ofApfel risk factors (Table 2). Importantly, the incidenceof PONV varied from only 5% to 11% during the laterecovery period (24–72 h).

In assessing the percentage of patients who re-ported that PONV negatively impacted upon theirrecovery within the first 24 h after surgery, a greaterimpact was observed in the groups with higher riskscores (21 and 48 vs 3% in the 3 and 4 vs 2 risk factorgroups, respectively). In the late recovery period(24–72 h), there were no significant differences amongthe three risk factor groups with respect to any of theoutcome measures (Table 2).

DISCUSSIONThese prospective data suggest that PONV remains

a common problem for high-risk surgical populationsdespite the frequent use of two or more antiemeticdrugs for prophylaxis. Multimodal approaches in-volving the use of two or more prophylactic anti-emetic drugs, avoiding highly emetogenic anestheticsand analgesics, and insuring adequate hydration arestrongly recommended for all patients at increasedrisk of developing PONV.8–10 The use of multimodalstrategies has assumed increased importance in thecurrent health care environment in order to facilitate afast-track recovery after surgery.2

Apfel’s simplified risk scoring system11 has beenshown to provide clinically useful information forpredicting PONV within 24 h after surgery in patientsnot receiving prophylactic antiemetic drugs.12 Com-pared to predicting a patient’s risk for PONV based ona history of PONV or the type of surgery alone, the useof this simplified risk scoring system has been foundto be more sensitive and specific.13 In fact, a studyconfirmed that administering prophylactic antiemeticsto high-risk patients using this simplified scoring systemreduced institutional rates of PONV.14 However, the util-ity of the Apfel scoring system for predicting early(�24 h) and late (24–72 h) emetic symptoms has notbeen examined in patients receiving multimodal anti-emetic prophylaxis.

Despite the frequent use of prophylactic antiemet-ics in our at-risk study population, the Apfel scoringsystem was highly predictive of emetic symptomswithin 24 h after surgery. However, these data failedto find a correlation between the incidence of latePONV (�24 h after surgery) and patients’ Apfel riskscore. This finding may reflect the lower than expectedincidence of PONV symptoms during the 24–72 hpostoperative evaluation interval (due to the frequentuse of prophylactic antiemetics and rescue treatmentin the pre-discharge period). The incidence of “late”PONV in our study is less than the 35% incidencereported by Carroll et al.,15 but similar to the findingsin the systematic literature review by Wu et al.16

Another factor which may explain the lack of differ-ences among the three risk factor groups in the 24–72 hperiod is the potentially confounding effect of patientsself-administering oral opioid-containing analgesics inthe post-discharge period.10 It has been suggested thatthe use of oral opioid-containing analgesics may be asignificant contributing factor to postdischarge nauseaand vomiting.2 Post-discharge nausea and vomitingremains common after outpatient (and short-stay)surgery, and contributes to significant patient distressand can impair resumption of normal activities ofdaily living.15,16

Although more than 60% of the high-risk patients(with three or four risk factors) in our study receivedmultiple (�2) antiemetic medications for prophylaxis,most of these patients still required a rescue antiemeticin the postoperative period. More than 25% of thelower risk patients (with only two risk factors) alsorequired a rescue antiemetic even though that over80% of these patients received at least one prophylac-tic antiemetic drug. In contrast to the recommendationof Apfel et al.,13,14 these data suggest that even lowerrisk patient populations (e.g., those with two riskfactors) may benefit from multimodal antiemetic pro-phylaxis if they are undergoing major laparoscopic orplastic surgery procedures. Surprisingly, 24% of thepatients in the highest risk group received eithernothing or only one antiemetic drug for prophylaxis.

A major deficiency in this observational study wasrelated to the fact that the use of prophylactic anti-emetic therapy was left “to the discretion of theanesthesiologist.” However, this study was designedto evaluate current antiemetic practices in two higher-risk surgical populations, and to assess whether indi-vidual patient risk information provided by the Apfelrisk scoring system would have predictive value de-spite the frequent use of prophylactic antiemeticdrugs. A second deficiency of this analysis relates tothe fact that we only included patients with two ormore risk factors. However, we would agree withApfel et al.17 that patients with 0–1 risk factors are lesslikely to benefit from the routine use of prophylacticantiemetics. Finally, our failure to quantify theamount of perioperative fluid administered and thenumber of oral opioid-containing analgesics con-sumed in the post-discharge period are also deficien-cies in the study design.

Although prophylaxis per se does not alter thepredictive value of the Apfel risk criteria for earlyPONV, these data suggest that there remains a need todevelop improved antiemetic and anti-nausea thera-peutic regimens which are effective during both theearly postoperative period, as well as after the patienthas been discharged home. Although the use of allegedlylonger-acting antiemetics (e.g., aprepitant, palonose-tron, transdermal scopolamine)18–21 for prophylaxismight be capable of providing greater protectionagainst emetic symptoms in the post-discharge period,

462 PONV Risk Factors and Early vs Late Emesis ANESTHESIA & ANALGESIA

additional well-controlled studies are needed to sup-port their use in routine clinical practice.22

In conclusion, the Apfel risk scoring system waspredictive of PONV within 24 h after surgery even inthe presence of multimodal antiemetic prophylaxis.However, the Apfel risk score was not predictive ofpost-discharge emetic symptoms. The inability to dis-criminate among the various risk groups after 24 hmay be related to the low incidence of emetic symp-toms (5%–11% at 24–72 h), and the confounding effectof oral opioid analgesic medications in the post-discharge period. We conclude that pharmacologicalmanagement of PONV should be tailored to thepatients’ risk level using the Apfel scoring system tominimize the potential for adverse side effects in thepostoperative period.

REFERENCES

1. Watcha MF, White PF. Postoperative nausea and vomiting: itsetiology, treatment and prevention. Anesthesiology 1992;77:162–84

2. White PF, Kehlet H, Neal JM, Schricker T, Carr DB, Carli F,Fast-Track Surgery Study Group. The expanding role of anes-thesiology in fast-track surgery: from multimodal analgesia toperioperative medical care. Anesth Analg 2007;104:1380–96

3. Edler AA, Mariano ER, Golianu B, Kuan C, Pentcheva K. Ananalysis of factors influencing postanesthesia recovery afterpediatric ambulatory tonsillectomy and adenoidectomy. AnesthAnalg 2007;104:784–9

4. Tang J, Wang B, White, PF, Watcha MF, Qi J, Wender RH. Theeffect of timing on ondansetron administration on its efficacy,cost-effectiveness, and cost-benefit as a prophylactic antiemeticin ambulatory setting. Anesth Analg 1999;88:1191–2

5. Sadhasivam S, Saxena A, Kathirvel S, Kannan TR, Trikha A,Mohan V. The safety and efficacy of prophylactic ondansetronin patients undergoing radical mastectomy. Anesth Analg1999;89:1340–5

6. Sennaraj B, Shende D, Hadhasivam S, Ilavajady S, Jagan D.Management of post-strabismus nausea and vomiting in chil-dren using ondansetron: a value-based comparison of out-comes. Br J Anaesth 2002;89:473–8

7. White PF, Watcha MF. Postoperative nausea and vomiting;prophylaxis versus treatment. Anesth Analg 1999;89:1137–9

8. Scuderi PE, James RL, Harris L, Mims GR. Multimodal anti-emetic management prevents early postoperative vomiting afteroutpatient laparoscopy. Anesth Analg 2000;91:1408–14

9. Habib AS, White WD, Eubanks S, Pappas TN, Gan TJ. Arandomized comparison of multimodal management strategyversus combination antiemetics for the prevention of postopera-tive nausea and vomiting. Anesth Analg 2004;99:77–81

10. White PF. Prevention of postoperative nausea and vomiting- amultimodal solution to a persistent problem. N Engl J Med2004;350:2511–2

11. Apfel CC, Laara E, Koivuranta M, Greim CA, Roewer N. Asimplified risk score for predicting postoperative nausea andvomiting: conclusions from cross-validations between two cen-ters. Anesthesiology 1999;91:693–700

12. Pierre S, Benais H, Pouymayou J. Apfel’s simplified score mayfavourably predict the risk of postoperative nausea and vomit-ing. Can J Anaesth 2002;49:237–42

13. Apfel CC, Kranke P, Eberhart LHJ. Comparison of surgical siteand patient’s history with a simplified risk score for the predic-tion of postoperative nausea and vomiting. Anaesthesia2004;59:1078–82

14. Biedler A, Wermelt J, Kunitz O, Muller A, Wilhelm W, DethlingJ, Apfel CC. A risk adapted approach reduces the overallinstitutional incidence of postoperative nausea and vomiting.Can J Anaesth 2004;51:13–9

15. Carroll NV, Miederhoff P, Cox FM, Hirsch JD. Postoperativenausea and vomiting after discharge from outpatient surgerycenters. Anesth Analg 1995;80:903–9

16. Wu CL, Berenholtz SM, Pronovost PJ, Fleisher LA. Systematicreview and analysis of postdischarge symptoms after outpatientsurgery. Anesthesiology 2002;96:994–1003

17. Apfel CC, Bacher A, Biedler A, Danner K, Danzeisen O,Eberhart LH, Forst H, Fritz G, Hergert M, Frings G, Goebel A,Hopf HB, Kerger H, Kranke P, Lange M, Mertzlufft F, MotschJ, Paura A, Roewer N, Schneider E, Stoecklein K, Wermelt J,Zernak C. A factorial trial of six interventions for the preven-tion of postoperative nausea and vomiting. Anaesthesist2005;54:201–9

18. Tang J, D’Angelo R, White PF, Scuderi PE. The efficacy ofRS-25259, a long-acting selective 5-HT3 receptor antagonist, forpreventing postoperative nausea and vomiting after hysterec-tomy procedures. Anesth Analg 1998;87:462–7

19. Gesztesi Z, Scuderi PE, White PF, Wright W, Wender RH,D’Angelo R, Black LS, Dalby PL, MacLean D. Substance P(Neurokinin-1) antagonist prevents postoperative vomiting af-ter abdominal hysterectomy procedures. Anesthesiology2000;93:931–7

20. White PF, Tang J, Song D, Coleman JE, Wender RH, OgunnaikeB, Sloninsky A, Kapu R, Shah M, Webb T. Transdermal scopol-amine: an alternative to ondansetron and droperidol for theprevention of postoperative emetic symptoms. Anesth Analg2007;104:92–6

21. Gan TJ, Apfel CC, Kovac A, Philip BK, Singla N, Minkowitz H,Habib AS, Knighton J, Carides AD, Zhang H, Horgan KJ, EvansJK, Lawson FC, Aprepitant-PONV Study Group. A randomized,double-blind comparison of the NK1 antagonist, aprepitant,versus ondansetron for the prevention of postoperative nauseaand vomiting. Anesth Analg 2007;104:1082–9

22. Glass PS, White PF. Practice guidelines for the management ofpostoperative nausea and vomiting: past, present, and future.Anesth Analg 2007;105:1528–9

Vol. 107, No. 2, August 2008 © 2008 International Anesthesia Research Society 463