- Email: [email protected]
Opioid analgesia and assessment of the sonographic Murphy sign
The Journal of Emergency Medicine, Vol. 28, No. 4, pp. 409 – 413, 2005 Copyright © 2005 Elsevier Inc. Printed in the USA. All rights reserved 0736-4679/05 $–see front matter
doi:10.1016/j.jemermed.2004.12.009
Original Contributions
OPIOID ANALGESIA AND ASSESSMENT OF THE SONOGRAPHIC MURPHY SIGN Bret P. Nelson,
MD*,
Emily L. Senecal, MD*, Christine Hong†, Thomas Ptak, Stephen H. Thomas, MD, MPH*†
MD, PhD‡,
and
*Division of Emergency Medicine & Harvard Affiliated Emergency Medicine Residency, Harvard Medical School, Boston, Massachusetts, †Department of Emergency Services and ‡Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts Reprint Address: Stephen H. Thomas, MD, MPH, Department of Emergency Services, Massachusetts General Hospital, 55 Fruit Street, Clinics Building Room 115, Boston, MA 02114
e Abstract—Administration of intravenous opioid analgesia to patients with undifferentiated abdominal pain remains a controversial topic in many emergency departments. To determine whether opioid analgesia impacts assessment of the sonographic Murphy sign (SM) in evaluating acute gallbladder disease (GBD), a retrospective chart review was undertaken. The chart review encompassed 119 patients, 21% of whom, having received opioid analgesia before ultrasound, constituted the opioid group. Between the opioid and control (i.e., no opioid analgesia) groups, there were no significant differences in SM sensitivity (48.2%; CI 28.7– 68.1% vs. 68.8%; CI 41.3– 89%, respectively) or specificity (92.5%; CI 83.4 –97.5% vs. 88.9%; CI 51.8 –99.7%, respectively) for GBD. There was no association between opioid analgesia and false-positive SM (OR 0.74, CI 0.08 – 6.65), or false-negative SM (OR 1.42, CI 0.46 – 4.43). We conclude that the test characteristics of SM are unaffected by opioid analgesia. © 2005 Elsevier Inc.
RUQ pain exist, gallbladder disease (GBD) in the form of biliary colic and cholecystitis are among the most common, affecting roughly 25 million Americans every year and necessitating 600,000 choleycystectomies annually (1,2). The work-up of RUQ pain and the diagnosis of GBD depend heavily on the use of RUQ ultrasound (US), as the clinical and laboratory work-up for GBD has been shown to be inferior to US (3). Specific criteria evaluated by US include presence of gallstones, GB wall thickening, lumen distention, GB sludge, common bile duct dilatation, pericholecystic fluid, and the sonographic Murphy sign (SM). The sonographic Murphy sign is unique in that it is not merely an observed radiographic finding made by the emergency radiologist, but it relies on patient participation. The SM is defined as the presence of maximal tenderness occurring when the sonographer’s US transducer applies pressure over the visualized gallbladder. Several studies have evaluated the diagnostic parameters of the SM in the diagnosis of acute cholecystitis. Sensitivity and specificity ranges for the SM in acute cholecystitis have been reported as 63–96% and 35–98%, respectively (1,4 – 6). Likewise, the positive and negative predictive values have been reported as 43–73%, and 82-90%, respectively (5). When both the SM and gallstones are present on US, positive and negative predic-
e Keywords— oligoanalgesia; abdominal pain; ultrasound; sonographic Murphy sign; emergency department
INTRODUCTION Right upper quadrant (RUQ) pain is a chief complaint encountered regularly in emergency departments (EDs) throughout the country. Although numerous etiologies of
RECEIVED: 10 December 2003; FINAL ACCEPTED: 3 December 2004
SUBMISSION RECEIVED:
20 October 2004;
409
410
B. P. Nelson et al.
tive values for acute cholecystitis are 92% and 95%, respectively (7). Accuracy has been reported as 87–96% (4,6). Given the high frequency— up to 95% or higher (8)—with which SM is found in acute cholecystitis, the ability to detect the SM is a critical part of the diagnostic imaging workup for acute cholecystitis. During the work-up of undifferentiated abdominal pain (UAP) in the ED, emergency physicians (EPs) often administer opioid analgesia before obtaining a definitive diagnosis. The effect of opioid analgesia on the evaluation of UAP in the ED remains a controversial subject despite substantial literature supporting its judicious use (9 –16). Several articles have reviewed the existing body of literature on this topic, suggesting that it is probably safe to provide appropriate levels of pain relief, but a consensus has yet to be achieved (14,15,17). Given that the purpose of opioid analgesia is to diminish pain, and that eliciting the SM relies on the patient’s expression of pain, the question arises as to whether or not the administration of opioid analgesia before obtaining RUQ US affects the rate of positive SM in patients who “should” have this finding based upon their true diagnosis of GBD. There have been no studies examining this to date. A recent survey highlighted this controversy between emergency physicians and emergency radiologists. Although the emergency radiologists were unanimous (100%) in the belief that administration of intravenous opioid analgesia before RUQ US would negatively impact assessment of the SM, only 10% of emergency physicians agreed that such premedication would adversely affect the SM (18). We performed a retrospective cohort study as an initial step to assess effects of pre-US analgesia on the SM; this was done by examining SM test characteristics in patients who did (opioid group) and did not (control group) receive opioids before RUQ US. MATERIALS AND METHODS Study Design This project consisted of a retrospective cohort (chart review) study. The study was designed to test the null hypothesis that test characteristics of the SM are similar in patients who have, and in those who have not, received pre-US parenteral opioid analgesia. The study was reviewed by the Institutional Review Board.
annual census of about 75,000. The Attending staff is Emergency Medicine (EM) board-certified or eligible, and the ED radiology staff are board certified in Radiology, with many having added training in Emergency Radiology. During most of the day, attending emergency radiologists (all of whom have been at the study institution at least 10 years) did the US examinations. Between 11:00 a.m. and 7:00 a.m., senior radiology residents performed the US examinations. The study population consisted of 119 consecutive adult patients who underwent a RUQ US, ordered from the ED, during the study period of 2002. Measurements and Outcomes Medical records were reviewed for patient demographic data, timing and type of analgesia, presence of SM, radiographic diagnosis, and final ED diagnosis. The chart review also included assessment for any return ED visits occurring within 2 weeks of discharge from the index (initial) ED visit. This was performed to reduce chances of SM test characteristics bias, which would be risked if opioid administration were associated with falsenegative SM and ultimate ED diagnostic error (e.g., inappropriate discharge due to missed diagnosis related to false-negative US). As part of follow-up data collection, we assessed whether any patient discharged from the ED with a diagnosis of biliary colic returned to the ED within 2 weeks for the same illness. Also, pathology and operative reports were documented for patients undergoing cholecystectomy. The relevant null hypothesis of no intergroup (i.e., opioid vs. control) differences in SM performance was assessed using standard measures of test characteristics. Data Analysis Data were analyzed using STATA (Intercooled version 8 for Windows, StataCorp, College Station, TX). Analysis included chi-square and Fisher’s exact testing for categorical variables, the unpaired t-test for univariate assessments of continuous variables, and multivariate logistic regression with odds ratios (OR) and 95% confidence intervals (CI). Statistical significance was defined at the 0.05 level. RESULTS
Setting and Population
Baseline Characteristics
The study was conducted at a Level I, tertiary care ED that serves as a base for an EM residency and has an
In total, 119 charts were reviewed: 25 patients received opioid analgesia before RUQ US (opioid group), and 94
Analgesia and Sonographic Murphy Sign
411
Table 1. Baseline Characteristics in Patients who Received Opioids Before RUQ US and Those who Did Not Characteristic Mean age (SD) Female (%) Sonographic Murphy sign Gallbladder pathology (%)
Opioid (n ⫽ 25)
Control (n ⫽ 94)
p Value
51.2 (20) 16 (64) 12 (48)
48.3 (19) 56 (60) 18 (19)
0.49 0.76 0.003
16 (64)
26 (28)
0.001
did not (control group). Overall, the mean age of all study subjects was 49 years, and 61% were female. Baseline demographic characteristics between the two study groups were similar (Table 1). Cholecystitis or biliary colic was diagnosed in 43 patients (36%). The SM was positive in 30 patients (25%).
Analgesia Administration Of the 119 patients studied, 59 (49.6%) received analgesia of some type (primarily NSAIDS and/or opioids), and 25 (21%) received opioid analgesia before undergoing RUQ US. Thus, 60 patients (50.4%) received no analgesia before RUQ US.
Test Characteristics of the Sonographic Murphy Sign Between the opioid and control groups, there were no significant differences for SM in determining GBD with respect to either sensitivity (48.2%; 95% CI 28.7– 68.1% for opioid vs. 68.8%; 95% CI 41.3– 89% for control) or specificity (opioid 92.5%; 95% CI 83.4 –97.5% vs. control 88.9%; 95% CI 51.8 –99.7%) (Figure 1). Adjusting for patient age, sex and whether the US was performed outside of the “business hours” (11:00 p.m.–7:00 a.m.) of daytime/evening radiology attending coverage (US was performed by radiology residents otherwise), no association was found between the opioid pre-medication and either false-positive SM (OR 0.74, CI 0.08 – 6.65), or false-negative SM (OR 1.42, CI 0.46 – 4.43).
pathology, encouraging addiction, risking over-sedation, focus on diagnosis rather than treatment of painful conditions, and a lack of training in pain management techniques. Despite the fact that every patient in this study had abdominal pain of sufficient severity to warrant further investigation by RUQ US, only 49.6% of patients received any type of analgesia before their radiographic evaluation, and only 21% received opioid analgesia. Though it is possible that some patients had spontaneous relief of pain, it is also likely that our results on analgesia reflect suboptimal rates of medication administration. In fact, the possibility that even those who did receive analgesia had inadequate pain relief remains a potential confounder of our results: patients achieving little pain relief from opioids would not be expected to have alteration of the SM. Post-analgesia pain tracking, to document that the SM is unaltered with opioid administration effecting sufficient pain relief, will be an important part of any prospective trial addressing this issue. However, the data in the current study are relevant to the question, given findings that no significant opioid-related SM alteration seemed to occur over one year of standard clinical practice at a busy academic ED. EPs and surgeons have long contributed to the debate over opioid analgesia in the evaluation of undifferentiated abdominal pain (UAP). The surgical literature throughout the last century was replete with admonitions against masking physical examination findings with analgesia (23). However, during the last decade, many studies in the surgical and ED literature have more meticulously analyzed this matter and demonstrated the safety and efficacy of judicious analgesia in cases of UAP (9,12–14,16,17,24). Although the debate is far from resolved in everyday practice, ample evidence now exists that it is not only humane but also safe to treat pain before establishing a diagnosis. In fact, newer editions of “standard” abdominal pain surgical texts have remonstrated against denial of analgesia provision (24).
DISCUSSION Many studies have documented the longstanding problem of oligoanalgesia in the ED setting (19 –22). Undertreatment of painful conditions has been described in terms of long wait times before analgesia administration, or total lack of analgesia provision. Potential explanations for this phenomenon include fear of masking true
Figure 1. Sensitivity and specificity of the sonographic Murphy sign. Values are percentages, ⴞ 95% confidence intervals.
412
Today, however, we see the debate reemerging in the realm of emergency radiology. The arguments used by emergency radiologists cautioning against analgesia before assessing the SM mirror those that have historically been used by those concerned about masking abdominal examination findings. An overwhelming majority (90%) of EPs surveyed at our affiliated institutions felt that assessment of the SM would not be hindered by opioid analgesia, implying that literature attesting to the safety of analgesia before surgical evaluation would generalize to radiology as well. The fact that radiologists surveyed unanimously disagreed with this sentiment could reflect a lack of familiarity with literature outside the scope of their daily practice, or it could reflect appropriate concern (even in light of extant analgesia literature) over masking a sonographic finding that lends so much weight to the diagnosis of acute GBD. When the SM is present along with any of the other “hard” findings of cholecystitis (GB wall thickening, pericholecystic fluid, common bile duct dilatation, or gallstones), the accuracy of the test is approximately 87–96% in detecting cholecystitis (4,6). Thus, both radiologists and EPs must rely on the subjective portion of the test to a greater extent than is found in any other emergency radiographic study. Therefore, an assessment of whether the test characteristics of the SM would be influenced by opioid analgesia is important. The potential operator-dependence of the SM was the major reason we included in multivariate modeling, a variable denoting whether US was performed by attending or resident radiologists (this variable was non-significant in all analyses). The study’s sample size (in addition to its retrospective methodology) precludes definitive conclusions; the widths of our statistically calculated confidence intervals should be interpreted as indicative of imprecision. However, the results obtained over a 1-year period were not consistent with any profound effect of opioid pre-medication, upon the sensitivity or specificity of the SM in determining acute GBD. Multivariate analysis adjusting for patient age, sex and whether the US was performed after hours found no association between the opioid vs. control group status and false positive or false negative SM. There is potential for confounding clinical factors in our comparison of patients who received opioids, with those who did not. In that, we think that it is this problem, rather than study sample size, that is the main issue prompting a prospective trial. However, given that analgesia administration (or its lack) decisions were multifactorial (e.g., associated with patient, physician, and even nurse/caseload factors), and that some homogeneity of patients is implied by the fact that all were specifically getting RUQ US for GB disease rule-out, we believe that
B. P. Nelson et al.
the dichotomization of patients by opioids administration is a reasonable step for a preliminary report such as this one. The opioid group was comparable to the study group in that their demographics were similar, and all patients had RUQ US ordered to assess for gallbladder disease. In other respects (as noted in Table 1), the groups were different. We believe that the differences as outlined in Table 1 are consistent with the greater likelihood of clinicians’ administering opioids in patients with more severe pain, that was clinically judged as more likely GB disease. (The fact that RUQ US was ordered is consistent [at our institution] with a nontrivial pain severity; patients who present with little pain [and who have normal laboratory workup and examination] are usually discharged from the ED with outpatient follow-up.) Though the potential for confounding by other clinical parameters remains, we believe that the potential differences do not prevent the results from conveying a central message that analgesia does not affect the SM. For the purposes of this study, acute biliary colic and acute cholecystitis were combined to reflect the similar practical ramifications of the diseases. For example, many patients without a clinical or radiographic diagnosis of cholecystitis were admitted for urgent (not emergent) cholecystectomy based on persistent symptomatic biliary colic. Many others were discharged with close surgical follow-up, and returned for planned cholecystectomies. Thus, the radiographic findings of gallstones in combination with the patients’ overall clinical picture impacted their disposition and follow-up care. The increased frequency of GBD in the opioid versus the control group may reflect the likelihood that patients with acute GBD experienced greater pain, and therefore were more likely to receive opioid analgesia, than were patients who did not have acute GBD. Of note, this difference does not impact calculations of sensitivity and specificity, because these test characteristics are defined irrespective of the prevalence of disease in the population being studied. The “gold-standard” used in this study was the final ED diagnosis, as assigned by the emergency physicians. This diagnosis combined the results of all historical, physical examination, laboratory, and radiographic data, as well as the input from consultants. Follow-up on enrolled patients revealed that only four patients returned with complaints related to GBD within 2 weeks of their initial ED assessment. In three of the four cases, patients discharged with a diagnosis of biliary colic returned for persistent symptoms and underwent cholecystectomy with pathology reports indicating chronic choleycystitis. In these three cases, one had received opioid analgesia and two had not. The fourth patient, who had been diagnosed in the ED as having biliary colic, returned 2
Analgesia and Sonographic Murphy Sign
days later with recurrent symptoms and underwent a cholecystectomy revealing early acute cholecystitis. In this case, the patient had not received opioid analgesia in the ED. An ongoing research project at our institution will prospectively address the safety of opioid analgesia with respect to SM findings. With the prospective design we will be able to perform a more detailed assessment of ED pain control and follow-up, and the utilization of a double-blinded, randomized, controlled design will afford an improved means of addressing the question of analgesia and the SM. However, the results from this preliminary trial are not consistent with a significant risk of SM masking by opioid analgesia, and— despite indications from emergency radiologists to the contrary—we believe our preliminary results support a continued practice of judicious analgesia for patients with abdominal pain and suspected GBD. Hopefully, the issue of pain control will continue to be addressed in an effort to increase patient satisfaction while providing compassionate and accurate care.
413
8. 9. 10.
11. 12. 13.
14. 15.
16. 17.
REFERENCES 18. 1. Everhart JE, Khare M, Hill M, et al. Prevalence and ethnic differences in gallbladder disease in the United States. Gastroenterology 1999;117:632–9. 2. Vilgrain V, Palazzo L. Choledocholithiasis: role of US and endoscopic ultrasound. Abdom Imaging 2001;26:7–14. 3. Singer AJ, McCracken G, Henry MC, et al. Correlation among clinical, laboratory, and hepatobiliary scanning findings in patients with suspected acute cholecystitis. Ann Emerg Med 1996;28:267. 4. van Weelde BJ, Oudkerk M, Koch CW. Ultrasonography of acute cholecystitis: clinical and histological correlation. Diagn Imaging Clin Med 1986;55:190 –5. 5. Bree RL. Further observations on the usefulness of the sonographic Murphy sign in the evaluation of suspected acute cholecystitis. J Clin Ultrasound 1995;23:169 –72. 6. Ralls PW, Halls J, Lapin SA, Quinn MF, Morris UL, Boswell W. Prospective evaluation of the sonographic Murphy sign in suspected acute cholecystitis. J Clin Ultrasound 1982;10:113–5. 7. Ralls PW. Real-time sonography in suspected acute cholecystitis.
19. 20. 21. 22. 23. 24.
Prospective evaluation of primary and secondary signs. Radiology 1985;155:767–71. Simeone JF, Brink JA, Mueller PR, et al. The sonographic diagnosis of acute gangrenous cholecystitis: importance of the Murphy sign. AJR Am J Roentgenol 1989;152:289 –90. Attard AR, Corlett MJ, Kidner NJ, et al. Safety of early pain relief for acute abdominal pain. BMJ 1992;305:554 – 6. Garyfallou AG, O’Connor RE, Fulda GJ, Levine BJ. A controlled trial of fentanyl analgesia in emergency department patients with abdominal pain: can treatment obscure the diagnosis (abstract)? Acad Emerg Med 1997;4:424. LoVecchio F, Oster N, Sturmann K, et al. The use of analgesics in patients with acute abdominal pain. J Emerg Med 1997;15:775–9. Pace S, Burke TF. Intravenous morphine for early pain relief in patients with acute abdominal pain. Acad Emerg Med 1996;3: 1086 –92. Thomas ST, Silen W, Cheema F, et al. Effects of morphine analgesia on diagnostic accuracy in emergency department patients with abdominal pain: a prospective, randomized trial. J Am Coll Surg 2003;196:18 –31. Thomas ST, Silen W. Effect of diagnostic efficiency of analgesia for undifferentiated abdominal pain. Br J Surg 2003;90:5–9. Vermeulen B, Morabia A, Unger PF, et al. Acute appendicitis: influence of early pain relief on the accuracy of clinical and US findings in the decision to operate—a randomized trial. Radiology 1999;210:639 – 43. Zoltie N, Cust MP. Analgesia in the acute abdomen. Ann R Coll Surg Engl 1986;68:209 –10. Nissman SA, Kaplan LJ, Mann BD. Critically reappraising the literature-driven practice of analgesia administration for acute abdominal pain in the emergency room prior to surgical evaluation. Am J Surg 2003;4:291– 6. Nelson BP, Hong C, Ptak T, Thomas SH. Administration of opioid analgesia prior to right upper quadrant ultrasound does not affect assessment of the sonographic Murphy’s sign (Oral presentation at EM Congress international meeting, Sitges, Spain, September 15, 2003). Cordell W, Keene K, Giles B, Jones JB, Jones JH, Brizendine E. The high prevalence of pain in emergency department care. Am J Emerg Med 2002;20:165–9. Martin JJ, Moore GP. Pearls, pitfalls, and updates for pain management. Emerg Med Clin North Am 1997;15:399 – 413. Ngai B, Ducharme J. Documented use of analgesics in the emergency department and upon release of patients with extremity fractures. Acad Emerg Med 1997;4:1176 – 8. Wilson J, Pendleton J. Oligoanalgesia in the emergency department. Am J Emerg Med 1989;7:620 –3. Cope Z. The early diagnosis of the acute abdomen. New York, NY: Oxford University Press; 1921. Silen W. Cope’s early diagnosis of the acute abdomen, 20th edn. New York, NY: Oxford University Press; 2000.
doi:10.1016/j.jemermed.2004.12.009
Original Contributions
OPIOID ANALGESIA AND ASSESSMENT OF THE SONOGRAPHIC MURPHY SIGN Bret P. Nelson,
MD*,
Emily L. Senecal, MD*, Christine Hong†, Thomas Ptak, Stephen H. Thomas, MD, MPH*†
MD, PhD‡,
and
*Division of Emergency Medicine & Harvard Affiliated Emergency Medicine Residency, Harvard Medical School, Boston, Massachusetts, †Department of Emergency Services and ‡Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts Reprint Address: Stephen H. Thomas, MD, MPH, Department of Emergency Services, Massachusetts General Hospital, 55 Fruit Street, Clinics Building Room 115, Boston, MA 02114
e Abstract—Administration of intravenous opioid analgesia to patients with undifferentiated abdominal pain remains a controversial topic in many emergency departments. To determine whether opioid analgesia impacts assessment of the sonographic Murphy sign (SM) in evaluating acute gallbladder disease (GBD), a retrospective chart review was undertaken. The chart review encompassed 119 patients, 21% of whom, having received opioid analgesia before ultrasound, constituted the opioid group. Between the opioid and control (i.e., no opioid analgesia) groups, there were no significant differences in SM sensitivity (48.2%; CI 28.7– 68.1% vs. 68.8%; CI 41.3– 89%, respectively) or specificity (92.5%; CI 83.4 –97.5% vs. 88.9%; CI 51.8 –99.7%, respectively) for GBD. There was no association between opioid analgesia and false-positive SM (OR 0.74, CI 0.08 – 6.65), or false-negative SM (OR 1.42, CI 0.46 – 4.43). We conclude that the test characteristics of SM are unaffected by opioid analgesia. © 2005 Elsevier Inc.
RUQ pain exist, gallbladder disease (GBD) in the form of biliary colic and cholecystitis are among the most common, affecting roughly 25 million Americans every year and necessitating 600,000 choleycystectomies annually (1,2). The work-up of RUQ pain and the diagnosis of GBD depend heavily on the use of RUQ ultrasound (US), as the clinical and laboratory work-up for GBD has been shown to be inferior to US (3). Specific criteria evaluated by US include presence of gallstones, GB wall thickening, lumen distention, GB sludge, common bile duct dilatation, pericholecystic fluid, and the sonographic Murphy sign (SM). The sonographic Murphy sign is unique in that it is not merely an observed radiographic finding made by the emergency radiologist, but it relies on patient participation. The SM is defined as the presence of maximal tenderness occurring when the sonographer’s US transducer applies pressure over the visualized gallbladder. Several studies have evaluated the diagnostic parameters of the SM in the diagnosis of acute cholecystitis. Sensitivity and specificity ranges for the SM in acute cholecystitis have been reported as 63–96% and 35–98%, respectively (1,4 – 6). Likewise, the positive and negative predictive values have been reported as 43–73%, and 82-90%, respectively (5). When both the SM and gallstones are present on US, positive and negative predic-
e Keywords— oligoanalgesia; abdominal pain; ultrasound; sonographic Murphy sign; emergency department
INTRODUCTION Right upper quadrant (RUQ) pain is a chief complaint encountered regularly in emergency departments (EDs) throughout the country. Although numerous etiologies of
RECEIVED: 10 December 2003; FINAL ACCEPTED: 3 December 2004
SUBMISSION RECEIVED:
20 October 2004;
409
410
B. P. Nelson et al.
tive values for acute cholecystitis are 92% and 95%, respectively (7). Accuracy has been reported as 87–96% (4,6). Given the high frequency— up to 95% or higher (8)—with which SM is found in acute cholecystitis, the ability to detect the SM is a critical part of the diagnostic imaging workup for acute cholecystitis. During the work-up of undifferentiated abdominal pain (UAP) in the ED, emergency physicians (EPs) often administer opioid analgesia before obtaining a definitive diagnosis. The effect of opioid analgesia on the evaluation of UAP in the ED remains a controversial subject despite substantial literature supporting its judicious use (9 –16). Several articles have reviewed the existing body of literature on this topic, suggesting that it is probably safe to provide appropriate levels of pain relief, but a consensus has yet to be achieved (14,15,17). Given that the purpose of opioid analgesia is to diminish pain, and that eliciting the SM relies on the patient’s expression of pain, the question arises as to whether or not the administration of opioid analgesia before obtaining RUQ US affects the rate of positive SM in patients who “should” have this finding based upon their true diagnosis of GBD. There have been no studies examining this to date. A recent survey highlighted this controversy between emergency physicians and emergency radiologists. Although the emergency radiologists were unanimous (100%) in the belief that administration of intravenous opioid analgesia before RUQ US would negatively impact assessment of the SM, only 10% of emergency physicians agreed that such premedication would adversely affect the SM (18). We performed a retrospective cohort study as an initial step to assess effects of pre-US analgesia on the SM; this was done by examining SM test characteristics in patients who did (opioid group) and did not (control group) receive opioids before RUQ US. MATERIALS AND METHODS Study Design This project consisted of a retrospective cohort (chart review) study. The study was designed to test the null hypothesis that test characteristics of the SM are similar in patients who have, and in those who have not, received pre-US parenteral opioid analgesia. The study was reviewed by the Institutional Review Board.
annual census of about 75,000. The Attending staff is Emergency Medicine (EM) board-certified or eligible, and the ED radiology staff are board certified in Radiology, with many having added training in Emergency Radiology. During most of the day, attending emergency radiologists (all of whom have been at the study institution at least 10 years) did the US examinations. Between 11:00 a.m. and 7:00 a.m., senior radiology residents performed the US examinations. The study population consisted of 119 consecutive adult patients who underwent a RUQ US, ordered from the ED, during the study period of 2002. Measurements and Outcomes Medical records were reviewed for patient demographic data, timing and type of analgesia, presence of SM, radiographic diagnosis, and final ED diagnosis. The chart review also included assessment for any return ED visits occurring within 2 weeks of discharge from the index (initial) ED visit. This was performed to reduce chances of SM test characteristics bias, which would be risked if opioid administration were associated with falsenegative SM and ultimate ED diagnostic error (e.g., inappropriate discharge due to missed diagnosis related to false-negative US). As part of follow-up data collection, we assessed whether any patient discharged from the ED with a diagnosis of biliary colic returned to the ED within 2 weeks for the same illness. Also, pathology and operative reports were documented for patients undergoing cholecystectomy. The relevant null hypothesis of no intergroup (i.e., opioid vs. control) differences in SM performance was assessed using standard measures of test characteristics. Data Analysis Data were analyzed using STATA (Intercooled version 8 for Windows, StataCorp, College Station, TX). Analysis included chi-square and Fisher’s exact testing for categorical variables, the unpaired t-test for univariate assessments of continuous variables, and multivariate logistic regression with odds ratios (OR) and 95% confidence intervals (CI). Statistical significance was defined at the 0.05 level. RESULTS
Setting and Population
Baseline Characteristics
The study was conducted at a Level I, tertiary care ED that serves as a base for an EM residency and has an
In total, 119 charts were reviewed: 25 patients received opioid analgesia before RUQ US (opioid group), and 94
Analgesia and Sonographic Murphy Sign
411
Table 1. Baseline Characteristics in Patients who Received Opioids Before RUQ US and Those who Did Not Characteristic Mean age (SD) Female (%) Sonographic Murphy sign Gallbladder pathology (%)
Opioid (n ⫽ 25)
Control (n ⫽ 94)
p Value
51.2 (20) 16 (64) 12 (48)
48.3 (19) 56 (60) 18 (19)
0.49 0.76 0.003
16 (64)
26 (28)
0.001
did not (control group). Overall, the mean age of all study subjects was 49 years, and 61% were female. Baseline demographic characteristics between the two study groups were similar (Table 1). Cholecystitis or biliary colic was diagnosed in 43 patients (36%). The SM was positive in 30 patients (25%).
Analgesia Administration Of the 119 patients studied, 59 (49.6%) received analgesia of some type (primarily NSAIDS and/or opioids), and 25 (21%) received opioid analgesia before undergoing RUQ US. Thus, 60 patients (50.4%) received no analgesia before RUQ US.
Test Characteristics of the Sonographic Murphy Sign Between the opioid and control groups, there were no significant differences for SM in determining GBD with respect to either sensitivity (48.2%; 95% CI 28.7– 68.1% for opioid vs. 68.8%; 95% CI 41.3– 89% for control) or specificity (opioid 92.5%; 95% CI 83.4 –97.5% vs. control 88.9%; 95% CI 51.8 –99.7%) (Figure 1). Adjusting for patient age, sex and whether the US was performed outside of the “business hours” (11:00 p.m.–7:00 a.m.) of daytime/evening radiology attending coverage (US was performed by radiology residents otherwise), no association was found between the opioid pre-medication and either false-positive SM (OR 0.74, CI 0.08 – 6.65), or false-negative SM (OR 1.42, CI 0.46 – 4.43).
pathology, encouraging addiction, risking over-sedation, focus on diagnosis rather than treatment of painful conditions, and a lack of training in pain management techniques. Despite the fact that every patient in this study had abdominal pain of sufficient severity to warrant further investigation by RUQ US, only 49.6% of patients received any type of analgesia before their radiographic evaluation, and only 21% received opioid analgesia. Though it is possible that some patients had spontaneous relief of pain, it is also likely that our results on analgesia reflect suboptimal rates of medication administration. In fact, the possibility that even those who did receive analgesia had inadequate pain relief remains a potential confounder of our results: patients achieving little pain relief from opioids would not be expected to have alteration of the SM. Post-analgesia pain tracking, to document that the SM is unaltered with opioid administration effecting sufficient pain relief, will be an important part of any prospective trial addressing this issue. However, the data in the current study are relevant to the question, given findings that no significant opioid-related SM alteration seemed to occur over one year of standard clinical practice at a busy academic ED. EPs and surgeons have long contributed to the debate over opioid analgesia in the evaluation of undifferentiated abdominal pain (UAP). The surgical literature throughout the last century was replete with admonitions against masking physical examination findings with analgesia (23). However, during the last decade, many studies in the surgical and ED literature have more meticulously analyzed this matter and demonstrated the safety and efficacy of judicious analgesia in cases of UAP (9,12–14,16,17,24). Although the debate is far from resolved in everyday practice, ample evidence now exists that it is not only humane but also safe to treat pain before establishing a diagnosis. In fact, newer editions of “standard” abdominal pain surgical texts have remonstrated against denial of analgesia provision (24).
DISCUSSION Many studies have documented the longstanding problem of oligoanalgesia in the ED setting (19 –22). Undertreatment of painful conditions has been described in terms of long wait times before analgesia administration, or total lack of analgesia provision. Potential explanations for this phenomenon include fear of masking true
Figure 1. Sensitivity and specificity of the sonographic Murphy sign. Values are percentages, ⴞ 95% confidence intervals.
412
Today, however, we see the debate reemerging in the realm of emergency radiology. The arguments used by emergency radiologists cautioning against analgesia before assessing the SM mirror those that have historically been used by those concerned about masking abdominal examination findings. An overwhelming majority (90%) of EPs surveyed at our affiliated institutions felt that assessment of the SM would not be hindered by opioid analgesia, implying that literature attesting to the safety of analgesia before surgical evaluation would generalize to radiology as well. The fact that radiologists surveyed unanimously disagreed with this sentiment could reflect a lack of familiarity with literature outside the scope of their daily practice, or it could reflect appropriate concern (even in light of extant analgesia literature) over masking a sonographic finding that lends so much weight to the diagnosis of acute GBD. When the SM is present along with any of the other “hard” findings of cholecystitis (GB wall thickening, pericholecystic fluid, common bile duct dilatation, or gallstones), the accuracy of the test is approximately 87–96% in detecting cholecystitis (4,6). Thus, both radiologists and EPs must rely on the subjective portion of the test to a greater extent than is found in any other emergency radiographic study. Therefore, an assessment of whether the test characteristics of the SM would be influenced by opioid analgesia is important. The potential operator-dependence of the SM was the major reason we included in multivariate modeling, a variable denoting whether US was performed by attending or resident radiologists (this variable was non-significant in all analyses). The study’s sample size (in addition to its retrospective methodology) precludes definitive conclusions; the widths of our statistically calculated confidence intervals should be interpreted as indicative of imprecision. However, the results obtained over a 1-year period were not consistent with any profound effect of opioid pre-medication, upon the sensitivity or specificity of the SM in determining acute GBD. Multivariate analysis adjusting for patient age, sex and whether the US was performed after hours found no association between the opioid vs. control group status and false positive or false negative SM. There is potential for confounding clinical factors in our comparison of patients who received opioids, with those who did not. In that, we think that it is this problem, rather than study sample size, that is the main issue prompting a prospective trial. However, given that analgesia administration (or its lack) decisions were multifactorial (e.g., associated with patient, physician, and even nurse/caseload factors), and that some homogeneity of patients is implied by the fact that all were specifically getting RUQ US for GB disease rule-out, we believe that
B. P. Nelson et al.
the dichotomization of patients by opioids administration is a reasonable step for a preliminary report such as this one. The opioid group was comparable to the study group in that their demographics were similar, and all patients had RUQ US ordered to assess for gallbladder disease. In other respects (as noted in Table 1), the groups were different. We believe that the differences as outlined in Table 1 are consistent with the greater likelihood of clinicians’ administering opioids in patients with more severe pain, that was clinically judged as more likely GB disease. (The fact that RUQ US was ordered is consistent [at our institution] with a nontrivial pain severity; patients who present with little pain [and who have normal laboratory workup and examination] are usually discharged from the ED with outpatient follow-up.) Though the potential for confounding by other clinical parameters remains, we believe that the potential differences do not prevent the results from conveying a central message that analgesia does not affect the SM. For the purposes of this study, acute biliary colic and acute cholecystitis were combined to reflect the similar practical ramifications of the diseases. For example, many patients without a clinical or radiographic diagnosis of cholecystitis were admitted for urgent (not emergent) cholecystectomy based on persistent symptomatic biliary colic. Many others were discharged with close surgical follow-up, and returned for planned cholecystectomies. Thus, the radiographic findings of gallstones in combination with the patients’ overall clinical picture impacted their disposition and follow-up care. The increased frequency of GBD in the opioid versus the control group may reflect the likelihood that patients with acute GBD experienced greater pain, and therefore were more likely to receive opioid analgesia, than were patients who did not have acute GBD. Of note, this difference does not impact calculations of sensitivity and specificity, because these test characteristics are defined irrespective of the prevalence of disease in the population being studied. The “gold-standard” used in this study was the final ED diagnosis, as assigned by the emergency physicians. This diagnosis combined the results of all historical, physical examination, laboratory, and radiographic data, as well as the input from consultants. Follow-up on enrolled patients revealed that only four patients returned with complaints related to GBD within 2 weeks of their initial ED assessment. In three of the four cases, patients discharged with a diagnosis of biliary colic returned for persistent symptoms and underwent cholecystectomy with pathology reports indicating chronic choleycystitis. In these three cases, one had received opioid analgesia and two had not. The fourth patient, who had been diagnosed in the ED as having biliary colic, returned 2
Analgesia and Sonographic Murphy Sign
days later with recurrent symptoms and underwent a cholecystectomy revealing early acute cholecystitis. In this case, the patient had not received opioid analgesia in the ED. An ongoing research project at our institution will prospectively address the safety of opioid analgesia with respect to SM findings. With the prospective design we will be able to perform a more detailed assessment of ED pain control and follow-up, and the utilization of a double-blinded, randomized, controlled design will afford an improved means of addressing the question of analgesia and the SM. However, the results from this preliminary trial are not consistent with a significant risk of SM masking by opioid analgesia, and— despite indications from emergency radiologists to the contrary—we believe our preliminary results support a continued practice of judicious analgesia for patients with abdominal pain and suspected GBD. Hopefully, the issue of pain control will continue to be addressed in an effort to increase patient satisfaction while providing compassionate and accurate care.
413
8. 9. 10.
11. 12. 13.
14. 15.
16. 17.
REFERENCES 18. 1. Everhart JE, Khare M, Hill M, et al. Prevalence and ethnic differences in gallbladder disease in the United States. Gastroenterology 1999;117:632–9. 2. Vilgrain V, Palazzo L. Choledocholithiasis: role of US and endoscopic ultrasound. Abdom Imaging 2001;26:7–14. 3. Singer AJ, McCracken G, Henry MC, et al. Correlation among clinical, laboratory, and hepatobiliary scanning findings in patients with suspected acute cholecystitis. Ann Emerg Med 1996;28:267. 4. van Weelde BJ, Oudkerk M, Koch CW. Ultrasonography of acute cholecystitis: clinical and histological correlation. Diagn Imaging Clin Med 1986;55:190 –5. 5. Bree RL. Further observations on the usefulness of the sonographic Murphy sign in the evaluation of suspected acute cholecystitis. J Clin Ultrasound 1995;23:169 –72. 6. Ralls PW, Halls J, Lapin SA, Quinn MF, Morris UL, Boswell W. Prospective evaluation of the sonographic Murphy sign in suspected acute cholecystitis. J Clin Ultrasound 1982;10:113–5. 7. Ralls PW. Real-time sonography in suspected acute cholecystitis.
19. 20. 21. 22. 23. 24.
Prospective evaluation of primary and secondary signs. Radiology 1985;155:767–71. Simeone JF, Brink JA, Mueller PR, et al. The sonographic diagnosis of acute gangrenous cholecystitis: importance of the Murphy sign. AJR Am J Roentgenol 1989;152:289 –90. Attard AR, Corlett MJ, Kidner NJ, et al. Safety of early pain relief for acute abdominal pain. BMJ 1992;305:554 – 6. Garyfallou AG, O’Connor RE, Fulda GJ, Levine BJ. A controlled trial of fentanyl analgesia in emergency department patients with abdominal pain: can treatment obscure the diagnosis (abstract)? Acad Emerg Med 1997;4:424. LoVecchio F, Oster N, Sturmann K, et al. The use of analgesics in patients with acute abdominal pain. J Emerg Med 1997;15:775–9. Pace S, Burke TF. Intravenous morphine for early pain relief in patients with acute abdominal pain. Acad Emerg Med 1996;3: 1086 –92. Thomas ST, Silen W, Cheema F, et al. Effects of morphine analgesia on diagnostic accuracy in emergency department patients with abdominal pain: a prospective, randomized trial. J Am Coll Surg 2003;196:18 –31. Thomas ST, Silen W. Effect of diagnostic efficiency of analgesia for undifferentiated abdominal pain. Br J Surg 2003;90:5–9. Vermeulen B, Morabia A, Unger PF, et al. Acute appendicitis: influence of early pain relief on the accuracy of clinical and US findings in the decision to operate—a randomized trial. Radiology 1999;210:639 – 43. Zoltie N, Cust MP. Analgesia in the acute abdomen. Ann R Coll Surg Engl 1986;68:209 –10. Nissman SA, Kaplan LJ, Mann BD. Critically reappraising the literature-driven practice of analgesia administration for acute abdominal pain in the emergency room prior to surgical evaluation. Am J Surg 2003;4:291– 6. Nelson BP, Hong C, Ptak T, Thomas SH. Administration of opioid analgesia prior to right upper quadrant ultrasound does not affect assessment of the sonographic Murphy’s sign (Oral presentation at EM Congress international meeting, Sitges, Spain, September 15, 2003). Cordell W, Keene K, Giles B, Jones JB, Jones JH, Brizendine E. The high prevalence of pain in emergency department care. Am J Emerg Med 2002;20:165–9. Martin JJ, Moore GP. Pearls, pitfalls, and updates for pain management. Emerg Med Clin North Am 1997;15:399 – 413. Ngai B, Ducharme J. Documented use of analgesics in the emergency department and upon release of patients with extremity fractures. Acad Emerg Med 1997;4:1176 – 8. Wilson J, Pendleton J. Oligoanalgesia in the emergency department. Am J Emerg Med 1989;7:620 –3. Cope Z. The early diagnosis of the acute abdomen. New York, NY: Oxford University Press; 1921. Silen W. Cope’s early diagnosis of the acute abdomen, 20th edn. New York, NY: Oxford University Press; 2000.