Radionuclide ventriculography in severely underweight anorexia nervosa patients before and during refeeding therapy

JOURNAL OF ADL\LESCENT I-K..%LTH 1991;12:301-,306

DANIEL

KAHN,

M.D.,

SCOTT B. PERLMAN,

JUANITA

HALLS,

M.D.,

J.A.

BIANCQ;

MI,D.,

AND

M.D.

Congestive heart failure is a well-recognized complication of refeeding therapy in underweight patients with anorexia nervosa but there are few data describing cardiac function during the critical refeeding period. This prospective study examined left ventricular function with conventional electrocardiographic-gated radionuclide ventricr~lograpby (RVG) in severely underweight anorexia nervosa patients both before and during refeeding therap]. Eight patients underwent rest and exercise RVG at admission and after regaining approx-: imately 5% to 10% of their ideal body weight. With the admission study serving as a control, the left ventricular ejection fraction and regional wall motion analysis were analyzed before and after refeeding and weight gain. Resting left ventricular ejection fractions were not sig nificantly different between the first and second RVGs (64 $z 11% vs. 62 f 8%6,respectively; P > .85). Likewise, the left ventricular ejection fraction with maximal exercise did not significantly differ when comparing the first or the second RVG (74 + 10% vs. 72 f 8%, P > .05). During the baseline RV,,p the left ventricular ejection fraction increased from 64 f 11% Crest) to 74 f 10% (maximal exercise) (P < .OOl).During the second RVG, the ejection fraction increased from 62 f 8% (rest) to 72 f 8% (maximal exercise) UJ = .003). Howevlrr, the left ventricular exercise ejection fraction in the second RVG in one patient increased only by one absolute percentage point. Four of the eight patients had regional wall motion

From the Department,; ofRadiology(D.K.; 1.A.B.; S.B.P.) and htternal Medicine (J.H.), University of Wisconsin Hospital and Clinics, Madison Wisconsin. Add&s reprint requests to: Daniel Kalm, M.D.. Department of Radiolog/, University of Iowa Hospitals and Clinics, Iowa Ciiy, IA 52242. Manuscript accepted December 27, 1990.

abnormalities detected during resting and/or exercise KVG. Abnormal cardiac function occurs in asymptomatic patients with anorexia nervosa undergoing refeeding therapy.

KEY WORDS:

Anorexia nervosa Cardiac Refeeding Radionuclide Vt3ltiicii~Ggi+p

The mortality rate of approximately 10% for patients with anorexia nervosa is higher than that in other psychiatric illnesses (l), and cardiac abnormalities are a common cause of death in these patients (2). Therefore, one of the major concerns for physicians is the risk of cardiac comphcations, especially during the refeeding stage of severely underweight patients. Congestive heart failure (3), as well as sudden death apparently from electrolyte disorders (4), has been reported during the nutritional rehabihtation period, though the pathophysiology has not been elucidated. Several studies have found abnormalities of cardiac morphology or function in patients with anorexia nervosa which include: ecnocardiogramdocumented reduction in cardiac mass and chamber dimensions (5-8), electrocardiographic abnormalities (4,9), decrease in cardiac output as measured by Swan-Ganz catheter (10) and echocardiogram (61 l), abnormal cardiovascular response to exercise with blunted heart rate and blood pressure recordings (5-

0 Society for Adolescent Meditine, 1991 _. Published by Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, NC-; fork, NY 10010

301 1054.139?u91/$3.50

302

JOURNAL OF ADOLkKiGii

KAHN ET AL.

7,12,13), and decreased maximum oxygen con(6,7,12}. However, in other reported studies, left ventricular systolic function as measured by echocardiogram (68) and electrocardiogram (ECG)gated radionuclide ventriculography (57) has been found to be normal. One study did find regional myocardial contraction abnormalities by echocardiogram in 8 of 14 patients with anorexia nervosa (11). Evaluation of cardiac dynamics before and during the early refeeding process is largely unstudied. This study, using rest and exercise ECG-gated radionuclide ventriculography (RVG), prospectively evaluated cardiac function in severely underweight patients with anorexia nervosa both before and during the refeeding period. sumption

.MeUzod.(: Patient Population We studied eight female patients, ages 18 to 36 years (mean f SD = 23.9 + 6.0 years). Selection criteria for study entrance were: 1) meet all the Diagnostic and Statisiiral Manual III-R criteria (14) for anorexia nervosa; 2) female, at Ieast I8 years of age; 3) admission weight less than or equal to 75% of ideal bcdy weight defined according to the 1959 Metropolitan weight tables (15); 4) expected length of hospital s&v at least 3 weeks. All patients admitted to the Eating Disorders llnit an the University of Wisconsin Hospital between April 1987 and April 1989 meeting the selection criteria were invited to participate. No patients had any history of heart failure. All patients were started on an 800-1200 calorie, noadded-salt diet and progressed by 300-400 calories every 2 to 4 days as tolerated to achieve a weight gain of 0.1 to 0.2 kg per day. Patients were switched to a regular diet 1 to 2 weeks after admission if no edema or signs of heart failure were present. The diet was supplemenied with Ensure if patients were unable to complete their meals. The research protocol was approved by the appropriate human subject committees at the University of Wisconsin Hospital, and informed consent was obtained in all patients.

:ZEALTH Vol. 12 No. 4

roid function) profiles were obtained in all patients prior to the first RVG, and when chnically indicated, prior to the second RVG. Diet and weight gain were carefully monitored. Patients underwent a baseline RVG within 6 days of admission (average 3.1 days). After the patient regained 5% to 10% of their ideal body weight, a second RVG was repeated. The RVG was performed in a standard fashion. Each patient received an intravenous injection of 0.7-1.0 mg of tin pyrophosphate. Twenty to 30 minutes later 10-20 mCi (370740 MBq) technetium-99m was intravenously administered and red blood cell labeling was accomplished. Anterior, ‘“best septal,” and left lateral views at rest were acquired on a large field of view scintillation camera ,(Elscint, Boston, MA). Immediately following the rest portion of the stud.y, the patient was exercised in the supine position with bicycle ergometry until 85% maximal heart rate was achieved, significant pain or fatigue occurred, or significant abnormalities were detected on simultaneous ECG or blood pressure recordings. Exercise was begun at a low level (25 W resistance) and increased by 25-watt intervals every 4 to 5 minutes to the maximum level. At yeak exercise, the same “best septal” view was acquired. Data Analysis A c9ntinuous loop tize of fJIP C&iac cyrle wwL gmerated from the ECG-gated acquisition, from which quantitative ejection fractions were calculated (16). Size, reejdnal wall motion analysis, and calculated left ~:*ntricular ejection fraction at rest and with exercise was interpreted independently by three experienced nuclear medicine specialists who were blinded to the details of the clinical histories or order of study. The resting myocardium was divided into seven segments and the “best septal” exercise image into three segments (Fig. 1) and interpreted as either

An!erlor

We*

808t Se~tsl (LAO) Waw

Left Latera

wew

1

c!z 3/f+

7

Data Acquisition AU patients underwent itdndard complete history and physical exar&iations, with a detailed review of their eatin disorder and cardiovascular history. Hem&logic and routine chemistry (including thy-

Figure 1. Anterior, best septal, and lefl lateral views obtained during the resting portions of the RVG’s; best septal view was obtained during the exerciseportion of the examination. Myocardial s2gmn! legend: 1 = anterobnsal, 2 = anterior, 3 = apex, 4 = septai, 5 = posterolaterai, 6 = inferior, 7 = inferokal.

June

1991

RADIONUCLIDE VENTRKULOGRAPHY

Table 1. Clinical characteristics of the patients,

--,

--

ideal

Percentage Admission ideal Prior Admission blood body Admission weight body low heart pressure Age weight Height (kg) weight at weight rate supine Patient (yr) x Frn) (ref. 15) admission (kg) (supine) (mm/Hg) (kg) A B C

19 21 27

34.4 41.8 42

152.4 165 163

51.7 59.5 58.3

66 70 72

37.6 43.1 40.8

60 50 96

102/70 100162 96R6

D

27

34.4

156

53.7

164

34.3

66

iPfii7b

E :

36

36.1

162

57.6

63

52

102l74

; H

_X 20 18

27.3 45 35.4

172 152 161.3

51.5 640 57.2

53 70 62

79 72 64

1lW70 IO/80 104!60

-29.5 27.3 36.7 35.9

303

Medications None None Nortriptyline Norethindrone/ mestranol Multivitamins premarin/provera None N;u&ibtyline Fiuoxetine Norethindroncl mestranol

Perceritagr ideal Caloric body Caloric 3:ieight intake intake RW ‘r regained RVG 2 ka?c,ries)by RVG 2 (calories) 8W 1500 1800

9 10

2200 2800 2500

1500

9

2800

1800

5

2800

1000 I800 1200

69 6

3200 3300 2200

’ Extrapolated as tables only go up to 169 cm.

normal, hypokinetic, akinetic, dyskinetic, or unable to evaluate. Size was interpreted as normal or enlarged (16). At least two of the three scan irlterpreters had to independently identify the same abnormality in chamber size or segflental wall motion for an abnormaltty to be documented. Statisticai Methods Paired sample Student’s t tests were usc:d to compare differences before and after weight gain for each patient P -C - .05 was considered significant. Results are reported as mean + standard deviation.

Results Clinical Charactedstirs (Tab% 1) The average age of onset of anorexia nervosa was 19.4 f 3.5 years and disease duration was 4.7 + 4.8 years. None of the patients had hypertension, diabetes, or had been diagnosed as having prior significant cardiac abnormalities. Patients F and H had abused laxatives, diuretics, or diet pills, but had stopped 1.5 to 2.0 years prior to this study. Though all the patients denied abuse of Syrup of Ipecac, patients D and H were vomiting {self-induced) daily up to the time of admission. Patient D drank 5 ounces of liquor per day; patient G had abused ethanol for 2 years up until 3 years prior to this stu+. None of the patients used cocaine. At the time of admission and Immediately prior to both the first and second RVG, patients B through

H had ho signs or symptoms of congestive heart failure; ie., no respiratory complaints, no edema, rales, or juguIar neck vein distension Only patient A had trace pedal edema which did not change over the C~DUFX of ‘her hospital admission. NO patients developed refeeding edema. Serum creatinine, thyroid function tests, total white blood cell count, and hematocrit were normal in all patients. Serum sodium, potassium and chloride were normal in all but two patients. Patient D had an admisslorl pU -tassium of 3.1 mEq/L which was corrected to 4.7 mEq/L by the time of the RVGs. Patient H had a serum potassium of 1.7 mEq/L which was corrected to 4.5 mEq/L by the time of the first RVG. In both cases, the hypokalemia was felt to be a result of vomiting. Apart from sinus bradyedrdia and U-waves associated with hypokalemia, there were no significant ECG findings. RVG Findings The patient’s weight at the time of the RVGs and the RVG data are presented in Fig. 2 and 3. Patients underwent a second RVG after regaining 5% to 10% of their ideal body weight 22 f 7 days (range lo32 days) after their first RVG. Mean patient weight was 37.6 + 5.8 kg at the time of the first RVG and increased to 41.1 -r- 5.7 kg by the time of the second RVG. Resting left ventlfctilar ejection fractions were not significantly different between the first and XCond RVGs (64 + 11% vs. 62 + a%, respectively). Likewise, the left ventricular ejection fraction with

304

!OURNAL OF ADOLESCENT HEALTH Vol. 12 No. 4

ICAHN ET AL.

30000 50 45 40

25000

95 SO 1-1 we!@1 oI2j 25 -i

20 I5

Rate Pressure PmdUCl 15000 (mmthpmlnj

10

RPPl

I

Is RPP2 __I

5 0

Figure 2. Weight at the time of admission and the jlrst RVG, and uw134tat the time ofthe second RVG.

maximal exercise did not significantly differ when comparing the first or the second RVG (74 + 10% vs. 72 & 8%). However, work performed during supine bicycle exercise, as reflected in the highest watt level sustained and maximal heart rate-systolic blood pressure product, did significantly increase between the first and second exercist: RVG (Fig. 4). The watt level increased from 56.2 + 22.2 W to 71.9

A

R

C

0

E

F

0

H

Svllpc

Fipra 3. Comprison of resting left ventric&r ejection +ctions fLVEF1 at ws; and during maximal exercisebefore(rest 1 and exercise 1) and after (rest 2 and exti::. ?! v+!z! @n.

0 A

e

c

D

E

F

a

H

Subject

Figura 4. Maximal work achieved estimated porn the heart rate-systoKc blood pressure product during first and second pressure product.

RVG’s. RPP = Rate-

f 16.0 W (P = ,011) and the rate-pressure product from 12,456 + 3879 mm Hg/min to 17,291 + 5211 mm Hg/min (P = -037). The mean left ventricular ejection fracticns normally and significantly increased during maximal exercise in both the first (baseline) and second (after weight gain) RVCs when compared with the resting ejection fractions. During the baseline RVG, the left ventricular ejection fraction increased tram 64 * 11% (rest) to 74 -e 10% (maximal exercise) (P < JOI). lhirtg the second RVG, the ejection fractions increased from 62 f 8% (rest) to 72 f 8% (maximal exercise) (P = .OO3).IIowever, in patient D, the left ventricular exercise ejection fraction in. the second RVG increased only by one absolute percentage point after an achieved rate-pressure product of 13,800 (6820 at rest). It must be noted, however, that the left ventricular ejection fraction in patient D increased from the baseline resting value of 49% to an ejection fraction at rest of 58% during the second RVG. In the remainder of the examinations, the increase in the exercise ejection fraction of the left ventricle was greater than 5 absolute percentage points in all patients.

June 1991

Four of the eight patients had regional wall IIIL\tion abnormalities detected during resting and/or exercise RVG. Patient A had a hypokinetic apical segment during the resting portion of the second RVG. l?atient D had an apical hypokinetic segment during the resting portion of the first RVG. Patient E had hypokinetic apical and inferior segments during the resting study of the first RVG. Patient F had hypokinetic apical and septal_segments during the rest portion of the first RVG with only the apical hypokinesis seen during the exercise portion of the first RVG. There were no akinetic or dyskinetic segments identified. All segments were ev,Fluable. In only one patient (F) was chamber Size felt to be abnormal: the left ventricle was mildly enlarged when viewed during the first, but not the secclnd; RVG. Discussion This study confirms some of the findings of other reports of cardiac studies in patients with anorexia nervosa, that is that the resting left ventricular ejection fraction is within normal limits. Our study design and entrance eligiKlity criteria differed frol;; other investigators (8,10,11) who have reported decreased cardiac output in patients wit& a:!orexia xxxvosa. Our series differs in that patients with bulimia were excluded trom our study, standard RVG was utilized instead of echocardiographically determined wall motion analysis and left ventricular ejection fraction, and patients served as their own controls before and durirlg oral refeeding therapy. Left ventricular response to exercise was evaluated before and during the early stages of the refeeding processes, as our study was specifically designed to investigate this critical period of early refeeding in this selected group of patients. In all but one patient we demonstrated that the left ventricular ejection fraction was normal both at rest and with exercise. This was seen in the very acute phase, when the body weight was less than 75% ideal body weight, and during the initial refeeding period of 2 to 4 weeks after admission, when the risk of heart faihire appears to be high (3). We did find regional wall motion abnormalities, primarily in the apical left ventricular region, with hypokinetic segments in four of eight patients. Two of these regions improved, one worsened, and one was unchanged during the refeeding period. Goldberg and co-workers (11) have found regional myocardial contraction abnormalities in 8 of 14 patients with anorexia nervosa by echocardiogram. While the clin-

RADIONUCLIDE VENTRICULOGRAI’Hy

305

ical significance of the findings related to wall motion is unclear, regional wall motion abnormalities in our series were observed in patients with anorexia nervosa who had no signs or symptoms of cardiac disease. Whether regional metabolic changes in the heart are induced by voluntary starvation and by the refeeding process reqires further study. Other modalities, like positron emission tomography, which can examine the regional myocardial bl:lod flow, QXygen extraction fraction, and glucose consumption, might provide important information concerning regional wall motion abnormalities in this group of patients. The main limitation of this study was the small number of subjects. While one patient demonstrated an abnormal left ventricular response to exercise, this is not certain. This patient had a significantly higher resting left ventricular ejection fraction during the second RVG when compared with the resting portion of the baseline resting RVG. It may be a normal response to demonstrate only a n&iiitial increase in t4c left ventricular ejection fraction when the resting ejection fraction is high (17). A larger series of patients would provide Ih’Loreconclusive data irr summary, our resuliz demonstrate that in this snail asrunptomatic group of patients with low body weight aving to anorexia nervosa, all have normal left ventricular ejection fractions at rest and at lea&t seve11of eight pstic?+ts cleady had a normal left ventricular response to exercise. One half of the patients demonstrated regional wall motion abnormalities. While it is p3ssiblc that the abnormalities detected might feyresent early subclinical heart failure, larger numbers of patients should be studied to assess this possibility. Further prospective studies are th,refore needed to define the cardiovascular dynamics during the critical early refeeding process in severely underweight patients with anorexia nervosa. This study was funded by the University of Wisconsin,Deparfment of Radiology Research and Development Fund.

References Task force on nomenclature and statistics of the An&can

fsychiatrir Association. Progress reports on the prepdration of DSM III. St. Louis, University of Missouri Press, 1976.

Silber T. Anorexia nervosa: Morbidity and mortality. Pediatr Ann 1984;13:851-9. Powers PS. Heart failure during treatment of anorexia nervosa. Am J Psychiatry 1982;139:1167-70. Isner JM, Roberts WC, Heymesfield SB, et al. Anorexia nervosa and sudden death. Ann Intern Med 1985;102:49-52.

306

KAHN ET AL.

5. Gottdiener JS, Gross HA, Henry WI ct al. Effects of selfstarvation on card&c size and function m anorexLa nervosa. Circulation 1978;58:425-33. 6. St John MG, Sutton M, Plapport T, et al. Effects of reduced left ventricular mass on chamber architecture, load and function: A study of anorexia nervosa. Circulation 1985;72:9911000. 7. Moodi DS, Sakozdo E. Cardiac function in adolescents and young adults with anorexia nervoea. J Adolesc Health Care 1983;49-14. 8. Heymesfield SB, Betherl R4, Ansley JD, et al. Cardiac abnornialities in cache&c patients before and during nutritional repletion. Am Heart J 1978;‘(5:5S4-94. 9. Thurston J, Mark P. Electrocardiographic abnormalities in patients with anorexia nervosa. Br Heart J ‘1974;36:719-23. 10. Kalager T, Brubakk 0, Bassoe HH. Cardiac performance in patients with anorexia nervosa. Cardiology 1978;63&4. 11. Goldbe?& SJ, Comerci GD, Feldman L. Cardiac output and regional myocardial contraction in anorexia nervosa. J Adolest Health Care 1988;9:15-21.

JOURNAL OF ADOLESCENT HEALTH Vol. 12 No. 4

12. Nude1 DB, Gootman N, Nussbaum MP, et al. Altered exercises performance and abnormal sympathetic responses to exercise in patients with anorexia nervosa. J Pediatr 1984;105:34-7. 13. Folin L, Freyschuss U, Bjarke B. Function and dimensions of the circulatory system in anorexia nervosa. Acta Paediatr Stand 1978;67:11-6. 14. American Psychiatric Association. Diagnostic and Stat$ical Manual of Mental Disorders, Third edition, Revised. W&hington DC, American Psychiatric Association, 19R7. 15. Blackbum GL, Bistrian BR, Maini BS, ~1 al. Nutitional and metabolic assessment of the hospitalized patient. JPEN 1977;1:11-22. 16. Schiepers C, Almasi JJ. Equilibrium g&ted blood pool imaging at test and during exercise. In: Gelfand MJ, Thomas SR, eds. Effective Use of Computers in Nuclear Medicine. New York, McGraw-Hill, 1988:136-60. 17. Adams RF, Vincent LM, Kimrey S, et al. Sex influences ventricular response to exercise in normals free of chest pain. Circulation 1985;72(111):425.