TO THE EDITOR:
We would like to comment on a flaw we observed in the methods of determining gallbladder ejection fraction using fatty-meal stimulation proposed by Krishnamurthy and Brown (1) and Jacobs and Peterson (2) in recent communications.
Both articles recognize the variable latent period between ingestion of a fatty meal and the onset of gallbladder contraction (B − A in Fig. 2 of Krishnamurthy and Brown (1)) but then propose that the ejection period used for ejection fraction calculations be determined at the trough of activity in the gallbladder or at the end of 60 min of imaging relative to peak activity (C − B in the same figure). In Figure 2 of Krishnamurthy and Brown (1), all 4 curves—w, x, y, and z—have variable ejection periods and have very different ejection fractions of approximately 20%, 40%, 60%, and 80%, respectively. However, the slope of gallbladder emptying is nearly identical for all 4 curves during the ejection phase, and therefore, gallbladder contraction is also nearly identical for all 4 curves even with the wide range of proposed ejection fractions (20%–80%).
A better estimate of gallbladder function would be the parameter that we use: Gallbladder ejection fraction is calculated using the ejection period between the onset of gallbladder contraction and 30 min after the onset of gallbladder contraction. The onset of gallbladder contraction is a marker of endogenous cholecystokinin secretion similar to exogenous Kinevac (sincalide for injection; Bracco Diagnostics, Inc.). A consistent 30 min of the ejection period is used for ejection fraction calculations instead of the variable-length ejection periods of the other methods. This simulates the 30-min-after-Kinevac gallbladder ejection fractions used for comparison with the normal values that clinicians use; that is, gallbladder ejection fraction is normal if it is >35%–40% at 30 min after injection of Kinevac.
In the past year, our clinicians have told us that our method produces gallbladder ejection fractions that help them discriminate between normal and abnormal gallbladder function. They continue to send patients to us for gallbladder ejection fraction determination to identify those with biliary dyskinesia and abdominal pain who would benefit from cholecystectomy or sphincterotomy. They also report that whether the patient’s usual symptoms were reproduced with fatty-meal ingestion is a clinically useful finding and should be included with the study results. We are now reviewing our results from 653 studies performed in 2002.
In order to use our proposed method of gallbladder ejection fraction determination, dynamic 1-min images for 60 min after ingestion of the fatty meal must be obtained to identify the onset of gallbladder contraction and determine the gallbladder ejection fraction. The protocol of Jacobs and Peterson (2) obtains images immediately after ingestion of the fatty meal and 15, 60, and 75 min after ingestion. This protocol would not have the temporal resolution necessary to accurately determine the onset of gallbladder contraction. Dynamic images are also necessary to differentiate between gallbladder contraction and filling patterns that represent cystic duct syndrome and those that represent dysfunction of the sphincter of Oddi (3).
Other differences in methods are the fatty meal used. Krishnamurthy and Brown (1) used 240 mL (8 oz) of half-and-half per 70 kg of body weight (24 g of fat and 1,339 kJ [320 kcal]). Jacobs and Peterson (2) used 90 mL (3 oz) of whipping cream with a teaspoon of sugar regardless of body weight (30 g of fat). We originally used Lipomul (Lee Pharmaceuticals) in the amount applied for gallbladder contraction studies with ultrasound protocols but had difficulties obtaining an adequate supply. We then switched to a cheaper alternative, Microlipid (Mead Johnson & Co.), that was readily obtainable through our hospital pharmacy. Microlipid has the same concentration of fat per milliliter as does Lipomul, so we used the same volume as Lipomul (volume in milliliters equals 0.68 × body weight in pounds or 1.5 × body weight in kilograms) for a standard fatty meal (105 mL per 70 kg of body weight equals 52.5 g of fat or 1,977 kJ [472.5 kcal]).
Advantages of the fatty meal are that it is much cheaper than a dose of Kinevac, avoids spasm of the sphincter of Oddi that can occur when Kinevac is injected too quickly, and truly mimics a patient’s meals. Disadvantages include less literature support for the use of fatty meals in nuclear medicine studies and a requirement for more gamma camera imaging time. In addition, some patients are nauseous and refuse to ingest the fatty meal or immediately regurgitate the ingested meal. Given our current excellent results with fatty-meal stimulation, we are debating whether to go back to Kinevac even now that it is available again. We may limit Kinevac use to only patients who cannot tolerate a fatty meal.
REPLY:
We appreciate the comments of Drs. Chen and Campbell on our article. As we tried to point out in our article, the value of gallbladder ejection fraction depends on various factors including fatty-meal content, total calories, and duration of postmeal data collection (1). Authors can choose to measure gallbladder ejection fraction during 30 min of ejection period as long as normal values are established for that fixed duration. Figure 2 of our article is only a schematic representation of various types of emptying curves one may encounter with half-and-half as the fatty meal at a dose rate of 240 mL (8 oz)/70 kg of body weight. Individual values are shown in Table 1. For patients, the highest gallbladder counts are normalized to 100% by the custom-designed hepatobiliary software we used (KHB-Quant). The slope of the curve may change for patients.
The responsibility for establishing normal values for a particular technique rests with the physician conducting and interpreting the study and not with the clinician who depends on the test result. Clinicians rely on the reference values cited next to the patient result. The physician who conducts the test and interprets the hepatobiliary study should validate the local technique, especially when it deviates from the published technique. Because it is expensive to perform studies on healthy subjects, and often difficult to justify ethically, patient outcome results are frequently used to establish normal limits (2).
While using the results of a published study, it is necessary to follow the borrowed technique in its entirety, including the content and calories of the total meal, and the dose rate and infusion duration of the octapeptide of exogenous cholecystokinin (CCK-8). Any deviation from the borrowed validated technique would call for revalidation. Gallbladder ejection fraction measured with the fatty meal ultimately depends on the total quantity of endogenous cholecystokinin released into the circulation. Disadvantages of the fatty meal are its longer test duration (2 h) and greater variability in ejection fraction (1). Because gallbladder emptying does not follow a gaussian distribution, one cannot rely on the traditional method of using the mean ± 2 SDs to set the normal range. An arbitrary lower limit needs to be chosen from the results of outcome studies (2). Gallbladder ejection fraction can be controlled to any desired level simply by varying the dose rate and duration of infusion of cholecystokinin (1–3). We use 35% as the lower limit of normal with a 3-min infusion of 10 ng CCK-8 per kilogram, and 50% as the lower limit with 3 ng/kg/min for 10 min (2,3). We hope that Drs. Chen and Campbell soon publish the results on their 653 patients to enable others to adopt their technique. Now that CCK-8 is again available for clinical use in the United States, the need for fatty-meal stimulation has lessened.
REPLY:
With the paucity of literature investigating the utility of fatty-meal gallbladder stimulation studies, we were pleased to read about Drs. Chen and Campbell’s experience and apparent success with their technique. Our fatty meal (90 mL [3 oz] of heavy whipping cream) has been well tolerated by our patients and was chosen because of our inability to obtain other standard meals previously described (such as Lipomul [Lee Pharmaceuticals] and Calogen [Scientific Hospital Supplies]) and the ability to deliver a high fat content in a small volume. Lactose intolerance has not been an issue, and for those unfortunate few with a “milk allergy,” we have given 30 mL (1 oz) of cooking oil with success.
Our technique for data acquisition has evolved such that only 2 static images are acquired: immediately before meal administration (after the initial dynamic 60-min acquisition) and 75 min later. Patient tolerability and the efficiency of gamma camera use were both improved by our not requiring the patient to lie motionless under the camera for more than 2 h. Like Drs. Chen and Campbell, we found that our referring physicians were well pleased with the results.
With the return of Kinevac (sincalide for injection; Bracco Diagnostics, Inc.), however, we have resumed doing our gallbladder stimulation studies using a 30-min infusion of Kinevac (0.02 μg/kg) and a 40-min dynamic acquisition. There continues to be more literature support for the use of Kinevac than for fatty-meal stimulation, and since many of our studies are done after hours, the overall time required for completion of the study is lessened. Perhaps the next time Kinevac is no longer available, our whipping cream may be served on a scoop of premium ice cream.