931 evidence of liver involvement since we found neither increases in the L, isoamylase previously associated with liver diseases6 nor any abnormality of liver function. The suggested specificity of an increased amylase/ creatinine clearance ratio in acute pancreatitis’ has been challenged because such an increase has also been found in diabetic ketoacidosis. 13 Because ketone bodies can interfere with creatinine assays on the ’AutoAnalyzer’,’14 the calculated clearance ratio may be spuriously high. If this pitfall is avoided, the clearance ratio should still be a valid index of pancreatitis in diabetic ketoacidosis. Salivary-type amylases are widely distributed in glandular epithelium-e.g., sweat glands,ts lacrymal glands, 16 human milk,17 prostatic secretions,18 ovary,!’" fallopian tube," cervical mucus,12 and lung.20 Indeed the serum-isoamylase patterns in the ketoacidotic patients are reminiscent of those in patients with diseases of genitalia, lung, and prostrate,12 and in other abnormal metabolic states such as postoperative hyperamylasæmia21 and even normal pregnancy.Z2 Phenformin therapy,23 which interferes with cellular aerobic metabolism, also increased serum-amylase, however, the isoenzyme type was not determined. If amylase were responsible for more than intraluminal gastrointestinal digestion and participated in intracellular metabolism, particularly in glycogen catabolism, hyperamylasaemia in diabetic ketoacidosis might reflect deranged carbohydrate metabolism in many tissues. Why hyperamylasæmia occurs in only 60% of these patients is not clear, but it is not apparently related to the degree of ketoacidosis or renal dysfunction. Our findings indicate that any pancreatic contribution to hyperamylasaemia must be very small in most cases of diabetic ketoacidosis. The importance of this observation was demonstrated in case 1, an 82-year-old woman with heart-disease and gallstones. The gallstones were thought to have precipitated pancreatitis and ketoacidosis, and she was prepared for cholecystectomy. However, isoenzyme analysis showed that the gallstones no
were not
of
responsible.
Requests for reprints should be addressed to A. L. W., Department Surgery, Massachusetts General Hospital, Boston, Massachusetts
02114, U.S.A. REFERENCES 1.
Knight, A. H., Williams, D. N., Ellis G., Goldberg,
D. M. Br.
med. J. 1973,
iii,128.
2. Tully, G. T., Lowenthal, J. J. Ann. intern. Med. 1958, 48, 310. 3. Belfiore, F., Napoli, E. Clin. Chem. 1973, 19, 387. 4. Salt, W. B., Schenker, S. Medicine, 1976, 55, 269. 5. Warshaw, A. L., Fuller, A. F. New Engl. J. Med. 1975, 292, 325. 6. Warshaw, A. L., Bellini, C. A., Lee, K. H. Gastroenterology, 1976, 70, 572. 7. Warshaw, A. L., Lee, K. H. ibid. 1976, 71, 388. 8. Hughes, P. D. Br. J. Surg. 1961, 49, 90. 9. Beigelman, P. M. Diabetes, 1971, 20, 490. 10. Finn, R., Cope, S. ibid. 1963, 12, 141. 11. Goldberg, D. M., Spooner, R. J., Knight, A. H. Clin. Chem. 1974, 20, 673. 12. Warshaw, A. L., Lee, K. H. J. surg. Res. (in the press). 13. Levine, R. I., Glauser, F. L., Berk, J. E. New Engl. J. Med. 1975, 292, 329. 14. Watkins, J. Clinica chim. Acta, 1967, 18, 191. 15. Levitt, M. D., Ellis, C., Engel, R. R. J. Lab. clin. med. (in the press). 16. VanHaeringen, N. J., Ensink, F., Glasius, E. Exp. Eye Res. 1975, 21, 395. 17. Fridhandler, L., Berk, J. E., Montgomery, K. A., Wong, D. Clin. Chem. 1974, 20, 547. 18. Gunaga, K., Sheth, A., Shanta, R. Indian J. med. Res. 1968, 56, 208. 19 Lehrner, L. M., Ward, J. C., Karn, R. C., Ehrlich, C. E., Merritt, A. D. Am. J. clin. Path. 1976. 20. Skude, G. Scand. J. Gastroent. 1975, 10, 577. 21 Morrissey, R., Berk, J. E., Fndhandler, L. Pelot, D. Ann. Surg. 1974, 180, 67. 22 Kaiser, R., Berk, J. E., Fridhandler, L. Am. J. Obstet. Gynec. 1975, 122, 283. 23. Williams, D. N., Knight, A. H., Goldberg, D. M. Postgrad med. J. 1974, 50, 765.
DOES GILBERT’S DISEASE EXIST? DAVID ROBINSON
ALAN BAILEY
BUPA Medical Centre, London N1
A. M. DAWSON St. Bartholomew’s Hospital, London EC1
18 454
and 5471 women attending centre had serum-bilirubin concentrations measured. 2.0% of men and 0.6% of women had concentrations of 25 µmol/l (1.5 mg/dl) or above. Both sexes showed skewed distributions for bilirubin concentration, but analysis of the data showed no evidence of bimodality for either sex. This suggests that people who are now diagnosed as having Gilbert’s disease may constitute the upper end of the normal range rather than a disease state. The term "constitutional hyperbilirubinæmia" is preferable, and its use can be coupled with reassurance to the patient that the condition causes no symptoms.
Summary
a
men
screening
Introduction A CHANCE finding of a raised serum-bilirubin on routine blood analysis or during investigation of symptoms is commonplace.’ The finding often causes alarm, to both patient and doctor, and leads to further tests. The eventual diagnosis is often Gilbert’s disease2 (Gilbert’s syndrome, constitutional hyperbilirubinaemia). The nature of this disease is in dispute, and many workers investigating people with Gilbert’s disease compare them with healthy controls,34 implying that they are abnormal. Clinical impression suggests that this is unsound. We have tried to find out if such people really are a separate group from normal by analysing data from a large number of people who have undergone a health screen.
Materials and Methods The sera of 18 454 men and 5471 women attending the BUPA Medical Centre for routine screening between March, 1974, and June, 1975, were analysed for bilirubin concentration,’ among other tests, on an SMA 12/60. 12 of the men with a serum-bilirubin of >25 µmol/l (1.55 mg/dl) who lived within easy reach of the centre were asked to attend for a further investigation. 7 responded to this invitation and were examined clinically, and their serum-bilirubin was estimated before and after a 36 h fast. After the fast the direct and indirect serum-bilirubin was estimated, a reticulocytecount was done, and the serum-alkaline-phosphatase and serumglutamic-oxaloacetic-transaminase (S.G.O.T.) were also repeated. The accuracy and quality control of the biochemical analyses have been reported previously.6 The data were entered onto 80-cotumn punched cards and analysed on the CDC 6500 computer at London University.
Results Both men (fig. 1)and women had a frequency distribution of serum-bilirubin with a skew to the right. The median bilirubin for men was 9.6 µmol/l which was significantly higher (p25 jjonol/1 who were recalled had the usual criteria of Gilbert’s syndrome-i.e., their serum-bilirubin rose when they fasted,11 13 the bilirubin was mainly unconjugated, and gross haemolysis was excluded. The minor haemolysis demonstrated in some patients with Gilbert’s syndrome is not enough to account for the hyperbilirubinaemia.14 It is not surprising that the repeat bilirubins were slightly different from the first ones. In patients with a
raised bilirubin the concentration does vary and the patients had not fasted before the second examination. Is Gilbert’s syndrome a disease or merely an extreme expression of normality? A low uridine-diphosphate glucuronyl transferase,’s altered bilirubin kinetics,4 slightly reduced red-blood-cell survival,14 or even a genetic predisposition3 do not mean that it represents a disease state. One would expect to find differences in factors which control bilirubin concentration in a group of people with a high bilirubin if they are compared with those with a lower bilirubin. Were it feasible to measure them in a large normal population these factors could also be expected to have a skew distribution. The frequency of symptoms in patients with Gilbert’s disease probably reflects the way they were diagnosed-i.e., by blood tests done to investigate vague lassitude or discomfort. Our recalled patients had no such symptoms. We suggest that "constitutional hyperbilirubinaemia" be used in preference to "Gilbert’s disease" so that the patient can be reassured that the condition produces no symptoms and so that unnecessary investigations will not be ordered during any future non-associated illness in which the bilirubin may be found to be further increased. Certainly, treatment with phenobarbitone’6 to lower the bilirubin and diminish vague symptoms seems
misplaced. Failure to recognise that the distribution of bilirubin concentration is skewed and the subsequent misapplication of gaussian statistical techniques can yield a misleadingly low upper limit of normal. An upper limit of 14 µmol/l (or 0-8 mg/dl) seems too low. The 98th percentiles in our population were 19 pLmol/1 (1-11 mg/dl) for females and 25 mol/1 (1-5 mg/dl) for males, and these values more realistically define the upper limit of "normality". Even on these strict criteria 2% of our screened men had abnormal bilirubin concentrations; almost all of them will have had no hepatocellular disease, but they
would, in the past, have been bert’s disease.
diagnosed
as
having
Requests for reprints should be addressed to D. R., Medical 210 Pentonville Road, London N1 9TA.
Gil-
Centre,
REFERENCES 1. Sherlock, S. Diseases of Liver and Biliary System; p. 249. Oxford, 1975. 2. Foulk, W. T., Butt, H. R., Owen, C. A., Whitcomb, F. F., Musin, H. L. Medicine, 1959, 38, 25. 3. Powell, L. W., Hemingway, E., Billing, B. H. Sherlock, S. New Engl. J. Med. 1967, 277, 1108. 4. Berk, P. D., Bloomer, J. R., Howe, R. B., Berlin, N. I. Am. J. Med. 1970,
49, 296. 5. Gambino, S. R. Stand. Methods clin. Chem. 1965, 5, 55. 6. Wilding, P., Rollason, J. G., Robinson, D. Clinica chim. Acta, 1972, 41, 375. 7. O’Kell, R. I. O., Elliott, R. Clin. Chem. 1970, 16, 161. 8. Warner, M., Tolls, R. E., Hultin, J. V., Melleckes, J. A. Klin. Chem. klin. Biochem. 1970, 8, 105. 9. Owens, D., Evans, J. J. med. Genet. 1975, 12, 152. 10. Roberts, L. B. Clinica Chim. Acta, 1967, 16, 69. 11. Elveback, L. R., Taylor, W. Ann. N. Y. Acad. Sci. 1969, 161, 538. 12. Felscher, B. I., Rickard, D., Redeker, A. G. New Engl. J. Med. 1970, 283, 170. 13. Owens, D., Sherlock, S. Br. med. J. 1973, iii, 559. 14. Powell, L. W., Billing, B. H., Williams, H. S. Aust. Ann. Med. 1967, 16, 221.
15. 16.
Black, M., Billing, B. H. New Engl. J. Med. 1969, 280, 1266. Black, M., Sherlock, S. Lancet, 1970, i, 1359.
RAISED URINARY FIBRIN-DEGRADATION
PRODUCTS, COMPLEMENT, AND IgG DURING AN INFLUENZA-LIKE ILLNESS L. FANANAPAZIR ELIZABETH EDMOND
MAUREEN ECCLESTON J. L. ANDERTON
Renal Unit and Department of Medicine, Western General Hospital, and Regional Virus Laboratory, City Hospital,
Edinburgh Urine from eight normal controls in whom an influenza-like illness developed contained high concentrations of fibrin-degradation products (F.D.P.), IgG, and C3. The study was carried out when influenza A was prevalent in the community. However, a wide range of serological investigations revealed no evidence for influenza A or other viruses. The infection may have been caused by other viruses which produce upper-respiratory-tract infections and which are not readily diagnosed by serology. Urinary fibrin-degradation products are a well-known marker of glomerulonephritic activity and viral antigens may have induced an immune-complex glomerulonephritis in the 8 controls in whom an influenza-like disease developed. A larger normal population should be investigated during a virus epidemic.
Summary
Introduction THE immunological basis for glomerulonephritis is well established and immunofluorescence studies have shown immunoglobulins in the glomeruli of patients with glomerulonephritis. Immunoglobulins are produced in response to an antigen, but the nature of the antigen is often obscure and its demonstration difficult. However, numerous bacterial antigens (especially streptococci) may cause glomerular damage through immunological mechanisms. Chronic viral infection of animals with lymphocytic choriomeningitis, the agent of aleutian dis-
were not
of
responsible.
Requests for reprints should be addressed to A. L. W., Department Surgery, Massachusetts General Hospital, Boston, Massachusetts
02114, U.S.A. REFERENCES 1.
Knight, A. H., Williams, D. N., Ellis G., Goldberg,
D. M. Br.
med. J. 1973,
iii,128.
2. Tully, G. T., Lowenthal, J. J. Ann. intern. Med. 1958, 48, 310. 3. Belfiore, F., Napoli, E. Clin. Chem. 1973, 19, 387. 4. Salt, W. B., Schenker, S. Medicine, 1976, 55, 269. 5. Warshaw, A. L., Fuller, A. F. New Engl. J. Med. 1975, 292, 325. 6. Warshaw, A. L., Bellini, C. A., Lee, K. H. Gastroenterology, 1976, 70, 572. 7. Warshaw, A. L., Lee, K. H. ibid. 1976, 71, 388. 8. Hughes, P. D. Br. J. Surg. 1961, 49, 90. 9. Beigelman, P. M. Diabetes, 1971, 20, 490. 10. Finn, R., Cope, S. ibid. 1963, 12, 141. 11. Goldberg, D. M., Spooner, R. J., Knight, A. H. Clin. Chem. 1974, 20, 673. 12. Warshaw, A. L., Lee, K. H. J. surg. Res. (in the press). 13. Levine, R. I., Glauser, F. L., Berk, J. E. New Engl. J. Med. 1975, 292, 329. 14. Watkins, J. Clinica chim. Acta, 1967, 18, 191. 15. Levitt, M. D., Ellis, C., Engel, R. R. J. Lab. clin. med. (in the press). 16. VanHaeringen, N. J., Ensink, F., Glasius, E. Exp. Eye Res. 1975, 21, 395. 17. Fridhandler, L., Berk, J. E., Montgomery, K. A., Wong, D. Clin. Chem. 1974, 20, 547. 18. Gunaga, K., Sheth, A., Shanta, R. Indian J. med. Res. 1968, 56, 208. 19 Lehrner, L. M., Ward, J. C., Karn, R. C., Ehrlich, C. E., Merritt, A. D. Am. J. clin. Path. 1976. 20. Skude, G. Scand. J. Gastroent. 1975, 10, 577. 21 Morrissey, R., Berk, J. E., Fndhandler, L. Pelot, D. Ann. Surg. 1974, 180, 67. 22 Kaiser, R., Berk, J. E., Fridhandler, L. Am. J. Obstet. Gynec. 1975, 122, 283. 23. Williams, D. N., Knight, A. H., Goldberg, D. M. Postgrad med. J. 1974, 50, 765.
DOES GILBERT’S DISEASE EXIST? DAVID ROBINSON
ALAN BAILEY
BUPA Medical Centre, London N1
A. M. DAWSON St. Bartholomew’s Hospital, London EC1
18 454
and 5471 women attending centre had serum-bilirubin concentrations measured. 2.0% of men and 0.6% of women had concentrations of 25 µmol/l (1.5 mg/dl) or above. Both sexes showed skewed distributions for bilirubin concentration, but analysis of the data showed no evidence of bimodality for either sex. This suggests that people who are now diagnosed as having Gilbert’s disease may constitute the upper end of the normal range rather than a disease state. The term "constitutional hyperbilirubinæmia" is preferable, and its use can be coupled with reassurance to the patient that the condition causes no symptoms.
Summary
a
men
screening
Introduction A CHANCE finding of a raised serum-bilirubin on routine blood analysis or during investigation of symptoms is commonplace.’ The finding often causes alarm, to both patient and doctor, and leads to further tests. The eventual diagnosis is often Gilbert’s disease2 (Gilbert’s syndrome, constitutional hyperbilirubinaemia). The nature of this disease is in dispute, and many workers investigating people with Gilbert’s disease compare them with healthy controls,34 implying that they are abnormal. Clinical impression suggests that this is unsound. We have tried to find out if such people really are a separate group from normal by analysing data from a large number of people who have undergone a health screen.
Materials and Methods The sera of 18 454 men and 5471 women attending the BUPA Medical Centre for routine screening between March, 1974, and June, 1975, were analysed for bilirubin concentration,’ among other tests, on an SMA 12/60. 12 of the men with a serum-bilirubin of >25 µmol/l (1.55 mg/dl) who lived within easy reach of the centre were asked to attend for a further investigation. 7 responded to this invitation and were examined clinically, and their serum-bilirubin was estimated before and after a 36 h fast. After the fast the direct and indirect serum-bilirubin was estimated, a reticulocytecount was done, and the serum-alkaline-phosphatase and serumglutamic-oxaloacetic-transaminase (S.G.O.T.) were also repeated. The accuracy and quality control of the biochemical analyses have been reported previously.6 The data were entered onto 80-cotumn punched cards and analysed on the CDC 6500 computer at London University.
Results Both men (fig. 1)and women had a frequency distribution of serum-bilirubin with a skew to the right. The median bilirubin for men was 9.6 µmol/l which was significantly higher (p25 jjonol/1 who were recalled had the usual criteria of Gilbert’s syndrome-i.e., their serum-bilirubin rose when they fasted,11 13 the bilirubin was mainly unconjugated, and gross haemolysis was excluded. The minor haemolysis demonstrated in some patients with Gilbert’s syndrome is not enough to account for the hyperbilirubinaemia.14 It is not surprising that the repeat bilirubins were slightly different from the first ones. In patients with a
raised bilirubin the concentration does vary and the patients had not fasted before the second examination. Is Gilbert’s syndrome a disease or merely an extreme expression of normality? A low uridine-diphosphate glucuronyl transferase,’s altered bilirubin kinetics,4 slightly reduced red-blood-cell survival,14 or even a genetic predisposition3 do not mean that it represents a disease state. One would expect to find differences in factors which control bilirubin concentration in a group of people with a high bilirubin if they are compared with those with a lower bilirubin. Were it feasible to measure them in a large normal population these factors could also be expected to have a skew distribution. The frequency of symptoms in patients with Gilbert’s disease probably reflects the way they were diagnosed-i.e., by blood tests done to investigate vague lassitude or discomfort. Our recalled patients had no such symptoms. We suggest that "constitutional hyperbilirubinaemia" be used in preference to "Gilbert’s disease" so that the patient can be reassured that the condition produces no symptoms and so that unnecessary investigations will not be ordered during any future non-associated illness in which the bilirubin may be found to be further increased. Certainly, treatment with phenobarbitone’6 to lower the bilirubin and diminish vague symptoms seems
misplaced. Failure to recognise that the distribution of bilirubin concentration is skewed and the subsequent misapplication of gaussian statistical techniques can yield a misleadingly low upper limit of normal. An upper limit of 14 µmol/l (or 0-8 mg/dl) seems too low. The 98th percentiles in our population were 19 pLmol/1 (1-11 mg/dl) for females and 25 mol/1 (1-5 mg/dl) for males, and these values more realistically define the upper limit of "normality". Even on these strict criteria 2% of our screened men had abnormal bilirubin concentrations; almost all of them will have had no hepatocellular disease, but they
would, in the past, have been bert’s disease.
diagnosed
as
having
Requests for reprints should be addressed to D. R., Medical 210 Pentonville Road, London N1 9TA.
Gil-
Centre,
REFERENCES 1. Sherlock, S. Diseases of Liver and Biliary System; p. 249. Oxford, 1975. 2. Foulk, W. T., Butt, H. R., Owen, C. A., Whitcomb, F. F., Musin, H. L. Medicine, 1959, 38, 25. 3. Powell, L. W., Hemingway, E., Billing, B. H. Sherlock, S. New Engl. J. Med. 1967, 277, 1108. 4. Berk, P. D., Bloomer, J. R., Howe, R. B., Berlin, N. I. Am. J. Med. 1970,
49, 296. 5. Gambino, S. R. Stand. Methods clin. Chem. 1965, 5, 55. 6. Wilding, P., Rollason, J. G., Robinson, D. Clinica chim. Acta, 1972, 41, 375. 7. O’Kell, R. I. O., Elliott, R. Clin. Chem. 1970, 16, 161. 8. Warner, M., Tolls, R. E., Hultin, J. V., Melleckes, J. A. Klin. Chem. klin. Biochem. 1970, 8, 105. 9. Owens, D., Evans, J. J. med. Genet. 1975, 12, 152. 10. Roberts, L. B. Clinica Chim. Acta, 1967, 16, 69. 11. Elveback, L. R., Taylor, W. Ann. N. Y. Acad. Sci. 1969, 161, 538. 12. Felscher, B. I., Rickard, D., Redeker, A. G. New Engl. J. Med. 1970, 283, 170. 13. Owens, D., Sherlock, S. Br. med. J. 1973, iii, 559. 14. Powell, L. W., Billing, B. H., Williams, H. S. Aust. Ann. Med. 1967, 16, 221.
15. 16.
Black, M., Billing, B. H. New Engl. J. Med. 1969, 280, 1266. Black, M., Sherlock, S. Lancet, 1970, i, 1359.
RAISED URINARY FIBRIN-DEGRADATION
PRODUCTS, COMPLEMENT, AND IgG DURING AN INFLUENZA-LIKE ILLNESS L. FANANAPAZIR ELIZABETH EDMOND
MAUREEN ECCLESTON J. L. ANDERTON
Renal Unit and Department of Medicine, Western General Hospital, and Regional Virus Laboratory, City Hospital,
Edinburgh Urine from eight normal controls in whom an influenza-like illness developed contained high concentrations of fibrin-degradation products (F.D.P.), IgG, and C3. The study was carried out when influenza A was prevalent in the community. However, a wide range of serological investigations revealed no evidence for influenza A or other viruses. The infection may have been caused by other viruses which produce upper-respiratory-tract infections and which are not readily diagnosed by serology. Urinary fibrin-degradation products are a well-known marker of glomerulonephritic activity and viral antigens may have induced an immune-complex glomerulonephritis in the 8 controls in whom an influenza-like disease developed. A larger normal population should be investigated during a virus epidemic.
Summary
Introduction THE immunological basis for glomerulonephritis is well established and immunofluorescence studies have shown immunoglobulins in the glomeruli of patients with glomerulonephritis. Immunoglobulins are produced in response to an antigen, but the nature of the antigen is often obscure and its demonstration difficult. However, numerous bacterial antigens (especially streptococci) may cause glomerular damage through immunological mechanisms. Chronic viral infection of animals with lymphocytic choriomeningitis, the agent of aleutian dis-
