A type 2 diabetic patient with liver dysfunction due to human insulin
Masato Tawata *, Masato Ikeda, Yasushi Kodama, Kaoru Aida, Toshimasa Onaya
Third Department ofInternal Medicine, Yamanashi Medical University, Tamaho, Yamanashi 409 3898, Japan
Received 18 October 1999; received in revised form 17 January 2000; accepted 31 January 2000

Diabetes Research and Clinical Practice, Vol 49/1 July 2000.

Keywords: Liver dysfunction; Cholestasis; Hepatocellular damages; Human insulin: Diabetes mellitus

1. Onset and course of the subject

A 45 year old male had been complaining of thirst and polydypsia for the last 3 months, and visited a private hospital. His height was 167 cm and weighted 57.8 kg. During about a month before the first visit to the private hospital, he had been restricting diet to some extent voluntarily based on a home medical book. He was diagnosed as having type 2 diabetes mellitus because his fasting plasma glucose showed 221 mg/dl with positive urinary ketone. He was hospitalized to the private hospital and instructed a diet of 1600 kcal. Until the admission, he had drank about 60g of ethanol every day for the last 10 years, and he was pointed out to have mild liver dysfunction 6 months before the admission. However, his liver function tests were within normal ranges except mild elevation in serum y-GTP (167 U/1, normal range below 62), and hepatitis related virus markers were negative (Table 1). Since his plasma glucose did not improve with positive urinary ketone, human insulin (Penfil 30R®) injection was started. His plasma glucose improved, and he was discharged on insulin. One month later, he was noticed to have elevated serum aminotransferases and hospitalized to the same hospital again. The values of y-GTP, ALT and AST peaked at 1938, 1284 and 1080 IU/1, respectively (Fig. I and Table 1 98/01/30). The elevation of y-GTP was dominant through the clinical course. In addition, LDIL ALP and total billrubin also increased (Table 1 98/01/30). Insulin injection was discontinued and serum y-GTP and aminotransferases and other liver function tests returned close to the normal range within 2 weeks. Since his glycemic control became poor again after the discontinuation of insulin, human insulin (Penfil 30R®') was restarted. Serum y-GTP and aminotransferases increased again after the reinitiation of insulin. Therefore, insulin was discontinued again and the patient was referred to the Third Department of Internal Medicine, Yamanashi Medical University because of liver dysfunction, and was hospitalized on February 24, 1998. The values of liver function tests still remained elevated (Fig. 1, arrow). He showed positive HCV antibody on admission to our hospital, and the liver function tests were all within normal ranges 4 weeks after the admission (Table 1 98/03/23). Ultrasound examination of liver showed mild to moderate fatty liver. During the mean time, his plasma glucose improved by diet therapy, and glycemic control became fairly good by the administration of glibenclamide.

However, since his urinary C-peptide was very low (7.6 µg/day, Table 1). after obtaining informed consent from the patient because of a possible use of insulin in the future. another human insulin preparation (Humalin R®) was challenged. Seven days after the insulin injection, serum aminotransferases began to elevate again (Fig. I and Table 1 98/03/30). The insulin injection was discontinued immediately. and liver function tests returned within normal ranges within 7 days.

Lymphocytes stimulation test (LST) revealed negative results against all human insulin preparations (Penfil R®. Penfil R®'. Humalin R®) examined. Throughout the course. the patient did not show any skin reactions, nor received oral drugs except glibenclamide.

Table I
Biochemical data of the case^a

	98/01/30	98/03/23	98/03/30
ALT [IU/I]	1080 (9-48)	15 (10-32)	44
AST [IU/I]	1284 (5-49)	16 (2-31)	86
LDH [IU//I]	521 (89-221)	127 (118-213)	158
Y-GTP [IU/11]	1938 (0-62)	92 (8-94)	82
ALP [IU/I]	458 (31-115)	305 (121-320)	327
T-Bil [mg/dl]	4.9 (0.2-1.0)	0.7 (0.3- 1.3)	0.6
HA IgM antibody	Negative
HBs antigen	Negative	Negative (on admission)
HCV antibody	Negative	>4096 (on admission)
HCV RNA	-	<0.5 mEq ml (on admission)
Insulin antibody	-	Negative (on admission)
LST	-	Negative
Urinary C-peptide		7.6 µg day

' Lymphocyte stimulation test (LST) was carried out against three human insulin preparations (Penfil R®. Penfil N® and Humalin R®. The values in square and round brackets indicate the units and the normal ranges, respecti\eK. HCV RNA was qpuntitated by bDNA probe method. The data of 98/01/30 were obtained in the previous hospital.

2. Discussion

Although we did not perform liver biopsy in this case, the data of liver function tests were dominant in the elevation of y-GTP compared to ALT or AST, which suggested that cholestatic process was more dominant than hepatocellular damages. Therefore, we speculate that acute hepantis due to HCV is not likely the cause of liver dysfunction in this case. Furthermore, although the patient showed positive HCV antibody, quantity of HCV RNA virus was less than detectable range. Moreover, although the patient showed fatty liver by ultrasound examination, insulin injection and withdrawal always synchronized with the increase and the decrease in serum y-GTP and aminotransferases levels. Therefore, we considered that virus related liver dysfunction or fatty liver were not at least a major cause in the present case but rather that the patient showed drug induced liver dysfunction due to human insulin preparations. We think that a negative result in LST study is compatible with drug induced liver dysfunction. Penfil® and Humalin® are biosynthetic human insulin preparations produced and purified from yeast and E. coli, respectively. Minor contaminants during the purification of human insulin have been reported in both preparations [1,2]. Since the elevation of serum aminotransferases was observed by different preparations of human insulin, we speculated that the patient showed liver dysfunction by human insulin itself rather than the minor contaminants. The peak of serum aminotransferases in the last episode was lower than the previous two episodes. However, we think it is because we discontinued insulin injection immediately after we recognized the elevation of serum aminotransferases by every day examination. We therefore concluded that human insulin itself caused liver dysfunction in this case. However, it is possible that fatty liver might have exaggerated insulin action on the elevation of aminotransferases and other liver function tests.

Most of the adverse reactions to insulin preparations include local erythematous, pururtic and indurated reactions at injection site or generalized urticaria or anaphylaxis [3,4]. In addition, many cases of diabetes mellitus with liver dysfunction in insulin treated patients have been reported so far. However, most of these cases were associated with diabetic ketoacidosis, fatty liver or glycogen storage [5-11]. So, it seems that diabetic ketoacidotic state of the patients or acute metabolic amelioration by insulin treatment seemed to be related to liver dysfunction rather than insulin itself. However, there are some cases whose liver dysfunction seemed to be due to insulin preparation. To the best of our knowledges, six such cases have been reported so far [12- 15] (Table 2). Among these cases, Chifu et al. [13], Tominaga et al. [14] and Kudo et al. [15] reported four cases with liver dysfunction, whose serum aminotransferases improved by the changes from human insulin to porcine insulin or another type of human insulin preparations. These observations suggest that minor contaminants in insulin preparations may have been responsible for liver dysfunction rather than human insulin itself [13-15]

On the other hand however. Hiramatsu et al. [12] reported two cases of liver dysfunction, whose serum aminotransferases did not improve by the change in insulin preparation. In the present case, liver dysfunction was observed irrespective of the human insulin preparations. No improvement in liver dysfunction by the change in human insulin preparation indicates that contaminants in human insulin preparations are not likely the cause of liver dysfunction. So like the cases reported by Hiramatsu et al. [12]. we speculate that human insulin itself was responsible for liver dysfunction in the present case.

In conclusion. the present case and the report by Hiramatsu et al. [12] suggest that we need to pav more attention to liver function tests when we start insulin inJection and bear in mind that human insulin itself could cause or exacerbate liver dysfunction especially in those with liver diseases.

Table 2

Authors	Maximum ALT/AST	Pathology	Change of insulin	LST	Reference
	(IU/1)		preparation
Hiramatsu et al.	400/700	Active chronic		positive	12
(1982)		hepatitis
	100/230	Non-specific	Not effective	positive	12
Chifu et al. (1988)	-	-	Effective	positive	13
	-	-	Effective	-	13
Tominaga et al.	326/458	Non-specific	Effective	Negative	14
(1990)
Kudo et al. (1998)	339/330	Non-specific	Effective	-	15
Present case	1080/1284	-	Not effective	Negative

LST, lymphocyte stimulation test. Effective and not effective in the column of 'chanLe of un,uhn rcparation' indicate that liver dysfunction improved and did not improve, respectively, by the change of insulin preparanon.

Acknowledgements

We are very grateful to Drs O.Takekawa and I.Takayama for referring the patient and providing the data during the hospitalization at their hospital.

References:

[1] J. Markussen, U. Damgaard, 1. Diers, et al., Biosynthesis of human insulin in yeast via single chain precursors, Diabetologia 29 (1986) 568A-569A.
[2) R.E. Chance, E.P. Kroeff, J.A. Hoffman, B.H. Frank Chemical, physical, and biological properties of biosynthetic human insulin, Diabetes Care 4 (1981) 147-154.
[3] K. Kosaka, K. Kuzuya, Y. Akanuma, H. Kajinuma, K. Matsuoka, M. Kikuchi, A multicenter clinical study of the premixed biosynthetic human insulin formulation, Kiso To Rinsho (in Japanese) 27 (1993) 2159-2170.
[4] R.D. DeShazo, J.A. Galloway, Insulin allergy, Ann. Al
lergy Asthma Immunol. 76 (1996) 217 218.
[5] H.D. Bronstein, P.A. Kantrowitz, F. Schsffner, Marked enlargement of the liver and transient ascites with the treatment of diabetic acidosis, New Engl. J. Med. 261 (1959) 1314 1318.
[6] T.W. Astin, Hepatic dysfunction during diabetic ketosis,
Br. J. Clin. Pract. 26 (1972) 134 136.
[71 W.G. Manderson. M.T. McKiddie, D.J. Manners, et al., Liver glycogen accumulation in unstable diabetes mellitus, Diabetes 17 (1968) 13 16.
[8] R. Fujimaki, S. Maeyania, T. Iwasaki, et al., A case of diabetic ketoacidosis histologically showing chronic nonsuppurative destructive cholangitis like abnormal bile duct with marked elevated biliary enzyme during insulin therapy, St. Marianna Med. Univ. J. (in Japanese) 17 (1989) 884 980.
[9] R. Olsson, C. Wesslau, T. William Olsson, L. Zettergren, Elevated aminotransferases and alkaline phosphatases in unstable diabetes mellitus without ketoacidosis or hypoglycemia, J. Clin. Gastroenterol. 11 (1989) 541 545.
[101 K.R. Falchuk, D. Conlin, The intestinal and liver complications of diabetes mellitus, Adv. Intern. Med. 38 (1993) 269 286.
[11] M. Nakamuta, M. Ohashi, K. Goto, Y. Tanabe, K. Hiroshige, H. Nawata, Diabetes mellitus associated glycogen storage hepatomegaly: report of a case and review of the Japanese literature, Fukuoka Igaku Zasshi 84 (1993) 354 558.
[12] K. Hirainatsu, S. Arimori, Two cases of diabetes mellitus with elevated tramsaminases by insulin treatment, J. Jpn. Diab. Soc. (in Japanese) 25 (1982) 1221 1228.
[13] T. Chifu, Y. Ueda, T. Shikaya, Two cases of drug induced liver dysfunction due to human insulin preparations, J. Jpn. Diab. Soc. (Abstract, in Japanese) 31 (1988) 441.
[14] K. Tominaga, T. Hashiguchi, S. Yonomiya, Y, Tanabe, K. Hiroshige, A case of diabetes mellitus with liver dysfunction during the treatment, Rinsho To Kenkyuu (in Japanese) 67 (1990) 201 204.
[15] T. Kudo, H. Shimizu, T. Irinoshima, M. Mori, A case of liver dysfunction after treatment of rapid insulin preparation, J. Jpn. Diab. Soc. (in Japanese) 41 (1998) 43 47.

© 2000 Elsevier Science Ireland Ltd. All rights reserved. With permission from Elsevier Science Ireland Ltd.