Hyperamylasemia in Bulimia Nervosa andHyperemesis Gravidarum

Carol Robertson* and Harry Millar

Royal Cornhill Hospital, Aberdeen, Scotland

Accepted 29 January 1998

Abstract: Objectives: The exact causes of hyperamylasemia detected in bulimia nervosa areunknown but it is presumed to be due either to repeated binging or to vomiting. This studyset out to investigate the importance of vomiting in producing the raised serum amylase andto clarify whether the amylase in pancreatic or salivary. Methods: Patients suffering fromhyperemesis gravidarum who were repeatedly vomiting in pregnancy but not binge eatinghad their total serum and pancreatic amylase measured. Bulimic patients and a controlsample of nonvomiting pregnant women were similarly studied. An assessment of the fre-quency and duration of vomiting and binging was also made. Results: Results show 45% (5)of bulimic patients had raised serum amylase, but none had a raised pancreatic amylase.Twenty-four percent (7) of the hyperemetic patients also had a raised serum amylase level, allwith a normal pancreatic amylase level. None of the nonvomiting pregnant patients had araised amylase. Discussion: Of patients with hyperemesis gravidarum who repeatedly vomitbut do not binge, a significant number had raised amylase. This suggests that it is the vomitingrather than the binge behavior that increases amylase in bulimic patients. This increasedamylase probably comes from the salivary gland. © 1999 by John Wiley & Sons, Inc. Int J EatDisord 26: 223–227, 1999.

Key words: bulimia nervosa; amylase; hyperemesis gravidarum; vomiting

INTRODUCTION

Since the original description of bulimia nervosa (Russell, 1979), researchers have con-firmed various biochemical abnormalities in bulimic patients (Mitchell, Pyle, Eckert, Hat-sukami, & Lentz, 1983), including raised serum amylase levels in 27.8% to 62% of patientsstudied (Mitchell et al., 1983; Jacobs & Schneider, 1985; Agras, Schneider, Arnow, Rae-burn, & Telch, 1989). Three studies of isoenzymes showed that it was the salivary ratherthan the pancreatic component that predominated (Kaplan, 1987; Humphries, Adams,Eckfeldt, Levitt, & McLlain, 1987; Kinzl, Biebl, & Herold, 1993). Kinzl et al. (1993) alsoperformed pancreatic ultrasound and established no abnormal pathology in those withraised amylase levels.

*Correspondence to: Dr. Carol Robertson, Senior Registrar, Royal Cornhill Hospital, Cornhill Road, Aberdeen,Scotland, AB25 2ZH.

© 1999 by John Wiley & Sons, Inc. CCC 0276-3478/99/020223-05

Prod. #1422

Walsh, Wong, Pesce, Hadigan, and Bodourian (1990) found that an increased amylaselevel was correlated with the frequency of binge eating and vomiting. Gwirtsman et al.(1989) established that the serum amylase level in bulimic patients fell following inpatienthospitalization, but increased when free passes from the hospital were allowed and fol-lowing a “controlled binge-purge.” Kinzl et al. (1993) found hyperamylasemia in 61% oftheir bulimics, but did not detect a raised level in obese binge eaters who did not vomit.Olmsted, Kaplan, and Rockert (1994) identified vomiting frequency as an important prog-nostic indicator in bulimia nervosa. A biological marker of this would be a useful clinicaltool in monitoring patients’ progress.

It remained to be clarified whether it is the binge or vomit behavior which increases theamylase levels in bulimic patients. It was therefore proposed to see if amylase is raised inpatients suffering from hyperemesis gravidarum who suffer frequent vomiting withoutbinging. There is no evidence that the serum amylase level is raised in the first trimesterof uncomplicated pregnancy (Kaiser, Berk, & Fridhandler, 1975; Naeije, Neuray, VanMelsen, & Delcourt, 1979; Stickland, Widish, & Perez, 1984). In one study of 4 patientsadmitted with severe vomiting in pregnancy, 3 had a raised total serum amylase level, butno comment was made about isoenzyme levels (DeVore, Bracken, & Berkowitz, 1980).

The aims of the study were to confirm whether there was a definite association ofhyperemesis gravidarum with raised serum amylase, to identify the origin of any raisedamylase (pancreas or salivary glands), and to compare the findings with two controlgroups, namely patients with uncomplicated pregnancy and bulimia. Ethical Committeeapproval was granted for this study and patients gave informed written consent to enterthe study.

METHODS

Patients admitted to the maternity hospital with a diagnosis of hyperemesis gravida-rum were interviewed by the first author and serum was obtained for analysis. Theinterviewer established demographic details, the frequency and duration of vomiting,pre-existing eating disorders, other complicating physical problems, and alcohol con-sumption. A control population of nonvomiting pregnant women of similar gestationwere similarly assessed, excluding any with symptoms of an eating disorder or othercomplicating physical condition. Female bulimic patients who were referred to the EatingDisorders Service were identified and serum was obtained for analysis. Their demo-graphic details, frequency and pattern of vomiting, and binging behavior were noted andtheir body mass index (BMI) was recorded.

Laboratory Investigation

The serum from the three groups of patients was analyzed using a Beckman SynchronCx System for measuring pancreatic and total serum amylase. No direct measurement ofserum salivary amylase was possible with the facilities available. However, in the absenceof major organ damage or disease, the difference between total and pancreatic amylase isassumed to be salivary.

RESULTS

Data were collected from 29 hyperemetic patients. None had a pre-pregnancy eatingdisorder, excessive levels of alcohol consumption, nor any pre-existing illness which

224 Robertson and Millar

could have altered the results. Ten were pregnant for the first time and three of thepregnancies resulted in twins. Eleven bulimic patients were studied and had a mean BMIof 22.7. Eight nonvomiting antenatal patients were also studied. Full results are shown inTable 1. At 95% confidence intervals, 10–44% of the hyperemetic patients and 17–77% ofthe bulimic patients would have a raised serum amylase level.

Among the hyperemesis of pregnancy patients, 2 of the 7 (29%) in the raised serumamylase group were carrying twin pregnancies, whereas only 1 of 22 (5%) of the normalamylase group had a multiple pregnancy. Amylase isoenzyme analysis identified nopatients who had a raised pancreatic amylase, indicating that the raised amylase level wasdue to a raised salivary component. None of the 8 antenatal control subjects suffered fromany pre-existing eating disorder nor consumed higher than average amounts of alcohol.All had normal serum amylase levels. There was the suggestion of a trend among thebulimics for the occurrence of vomiting and binging to be more frequent among thosewith a raised amylase level, but the numbers were too small to allow for formal statisticalcomparison.

DISCUSSION

The results show a previously unreported finding that in a substantial proportion ofhyperemetic patients, there was a raised total serum amylase but no rise in pancreaticamylase. This strongly indicates that the source of the increased amylase is the salivaryglands. Similar evidence of a rise in salivary amylase was also found in our bulimicpatients, supporting the findings in previous studies. As expected from previous studies,there was no rise in amylase in normal pregnancy uncomplicated by excessive vomiting.Therefore, the raised amylase in bulimia nervosa and hyperemesis is probably caused byvomiting.

Patients in the bulimia group had a much longer duration of illness (mean 49 months)than the hyperemetic group (mean 1.1 month) which is understandable given the natural

Table 1.

Bulimics(NormalAmylase)

Bulimics(AbnormalAmylase)

Hyperemetics(NormalAmylase)

Hyperemetics(AbnormalAmylase)

Antenatal(All)

Patients (% of group) 6 (55%) 5 (45%) 22 (76%) 7 (24%) 8 (100%)Average age (years) 24 23 29 29 28Mean duration of illness

(months) 78 41 1.16 0.9 *Average frequency of

vomiting per day 1.8 2.2 9.5 7.3 0Average frequency of

binging per week 5.8 8.8 0 0 0Average gestation

(weeks) — — 10.5 10 7.6Abnormal total amylase 0 5 0 7 0Mean total amylase 55 128 56 109 56Range of total amylasea 41–59 103–167 29–88 90–140 35–89Abnormal pancreatic

amylase 0 0 0 0 0

aLaboratory reference range–normal amylase <90 U/L.

Hyperamylasemia in Bulimia 225

history of both disorders. However, the pregnant patients did have a much higher fre-quency of vomiting than the bulimic patients who have a much more chronic stablepattern. These differences may account for the higher rates of hyperamylasemia in thebulimic group. It would be difficult to establish a group with a similar pattern of vomitingto the bulimic patients in whom a more serious chronic condition did not confoundinterpretation of the results.

Vomiting in bulimia nervosa, by repeatedly stimulating the salivary glands, may leadto increased amylase production and enlargement of salivary glands. An increase in thesize of the parotid glands is a well-recognized sign of bulimia nervosa. Although ingestionof high quantities of carbohydrate-rich food has also been suggested as a cause of raisedserum amylase (Gwirtsman et al., 1989), Dawes (1970) in his review of the literature foundlittle evidence that the salivary amylase concentration can be altered by variation ofdietary carbohydrate intake. Also, Kinzl et al. (1993) did not find an increase in serumamylase in obese binge eaters (who usually do not vomit).

Dutta, Douglass, Smalls, Nipper, and Levitt (1981) found raised salivary amylase in 10%of patients admitted for alcohol detoxification which may be due to the effects of alcoholon salivary tissue but some alcoholics also vomit frequently. Systemic diseases have beenassociated with a rise in salivary amylase (Berk & Findhandler, 1975; Warshaw & Lee,1976; Dutta et al., 1981), but these were major disorders seldom seen in the young femalepopulation who usually suffer from bulimia nervosa.

Humphries et al. (1987) and Kaplan (1987) warn against the risk of overinvestigating arise in serum amylase which may be of salivary origin. Heigh, Matz, Roberts, Steinberg,and Henry (1990) reported the case of a 28-year-old female whose vomiting, recurrentabdominal pain, and raised amylase level led to 14 hospital admissions before the amylasewas found to be of salivary origin and an eating disorder was diagnosed. We have recentexperience of one patient with a raised amylase level having inpatient investigation forabdominal pain without the doctors being aware of her eating disorder and of anotherpatient being advised about pancreatic disease when a raised amylase level (probably ofsalivary origin) was identified.

Studying larger numbers of patients will help to confirm our preliminary findings andreplicate the findings of others. In addition to measuring amylase in patients with bulimianervosa (who binge and usually vomit) and those with hyperemesis gravidarum (whovomit but do not binge), it will be informative to also further study obese binge eaters(who binge but usually do not vomit) to confirm whether the rise found in serum amylasein bulimia nervosa is wholly of salivary origin and whether it is wholly caused byvomiting. This will further clarify the diagnostic utility of amylase in eating disorders andform a basis for studying possible changes in amylase as patients improve.

The results of this study, while important to psychiatrists, are also relevant to the manyclinicians including physicians, surgeons, obstetricians, and general practitioners whomay identify hyperamylasemia in their patients and need to be aware of the likely asso-ciation between vomiting and raised salivary amylase levels.

The authors thank Dr. Heather Watson, Dr. Catriona Morrison, the general practitioners, and nursesof the Holburn Practice, Dr. Norman Smith, and Dr. John Eagles.

REFERENCES

Agras, W.S., Schneider, J.A., Arnow, B., Raeburn, S.D., & Telch, C.F. (1989). Cognitive-behavioural and response-prevention treatment for bulimia nervosa. Journal of Consulting and Clinical Psychology, 57 (2), 215–221.

226 Robertson and Millar

Berk, J. E., & Findhandler, L. (1975). Clinical application of amylase isoenzyme analysis. American Journal ofGastroenterology, 63, 457–463.

Dawes, C. (1970). Effects of diet on salivary secretion and composition. Journal of Dental Research Supplement,49 (6), 1263–1272.

Devore, G.R., Bracken, M., & Berkowitz, R.L. (1980). The amylase/creatinine clearance ratio in normal pregnancyand pregnancies complicated by pancreatitis, hyperemesis gravidarum and toxemia. American Journal ofObstetrics and Gynecology, 136 (6), 747–754.

Dutta, S.K., Douglass, W., Smalls, U.A., Nipper, H.C., & Levitt, M.D. (1981). Prevalence and nature of hyper-amylasemia in acute alcoholism. Digestive Diseases and Sciences, 26, 136–141.

Gwirtsman, H.E., Kaye, W.H., George, D.T., Carosella, N.W., Greene, R.C., & Jimerson, D.C. (1989). Hyperamy-lasemia and its relationship to binge-purge episodes: Development of a clinically relevant laboratory test.Journal of Clinical Psychiatry, 50, 196–204.

Heigh, R.I., Matz, J., Roberts, I.M., Steinberg, W.M., & Henry, J.P. (1990). Atypical eating disorder masqueradingas recurrent acute pancreatitis. Journal of Clinical Gastroenterology, 12 (1), 78–80.

Humphries, L.L., Adams, L.J., Eckfeldt, J.H., Levitt, M.D., & McLlain, C.J. (1987). Hyperamylasemia in patientswith eating disorders. Annals of Internal Medicine, 106, 50–52.

Jacobs, M.B., & Schneider, J.A. (1985). Medical complications of bulimia: A prospective evaluation. QuarterlyJournal of Medicine, 54 (214), 177–182.

Kaiser, R., Berk, J.E., Findhandler, L. (1975). Serum amylase changes during pregnancy. American Journal ofObstetrics and Gynecology, 122 (3), 283–286.

Kaplan, A.S. (1987). Hyperamylasemia and bulimia: A clinical review. International Journal of Eating Disorders,6 (4), 537–543.

Kinzl, J., Biebl, W., & Herold, M. (1993). Significance of vomiting for hyperamylasemia and sialadenosis inpatients with eating disorders. International Journal of Eating Disorders, 13 (1), 117–124.

Mitchell, J.E., Pyle, R.L., Eckert, E.D., Hatsukami, D., & Lentz, R. (1983). Electrolyte and other physiologicalabnormalities in patients with bulimia. Psychological Medicine, 13, 273–278.

Naeije, R., Neuray, F., Van Melsen, A., & Delcourt, A. (1979). Enhancement of the amylase-creatinine clearanceratio in pregnancy. Acta Obsteticia et Gynecologica Scandinavica, 58, 531–533.

Olmsted, M.P., Kaplan, A.S., & Rockert, W. (1994). Rate and prediction of relapse in bulimia nervosa. AmericanJournal of Psychiatry, 151, 738–743.

Russell, G. (1979). Bulimia nervosa: An ominous variant of anorexia nervosa. Psychological Medicine, 9, 429–448.Stickland, D.M., Widish, J., & Perez, R. (1984). Amylase and isoamylase activities in serum of pregnant women.

Obstetrics and Gynecology, 63 (3), 389–391.Walsh, B.T., Wong, L.M., Pesce, M.A., Hadigan, C.M., & Bodourian, S.H. (1990). Hyperamylasemia in bulimia

nervosa. Journal of Clinical Psychiatry, 51, 373–377.Warshaw, A.L., & Lee, K.-H. (1976). Characteristic alteration of serum isoenzymes of amylase in diseases of liver,

pancreas, salivary gland, lung and genitalia. Journal of Surgical Research, 22, 362–369.

Hyperamylasemia in Bulimia 227