283 These studies were supported by the Cholera Research Laboratory, which is sponsored by the governments of the People’s Republic of Bangladesh, U.S.A., U.K., and Australia, and by an International Center for Medical Research N.I.H. grant SR07AI10048-17. We thank Dr Robert Yolken of N.I.H., Bethesda, Maryland, for help in establishing the ELISA for rotavirus at the Cholera Research Laboratory, and P. K. B. Neogi, Shamsul Huda, Dr Abdullah Al-Mahmud, Dr Hamidur Rahman, and Janita Stevens for assistance. Reprint requests should be addressed to D.A.S., Cholera Research Laboratory, G.P.O. Box 128, Dacca-2, Bangladesh.

requiring 30C j units daily who did not have insulin antibodies and did not respond to prednisolone; the latter caused ketoacidosis, which was treated successfully by continuous low-dosage insulin infusion similar to that used by Page et al.2 Thereafter, insulin infusion in a dose as low as 50 units a day maintained adequate glucose homoeostasis. In view of this remarkable response, other patients with large insulin requirements were also given small doses of insulin intravenously.

REFERENCES

1. Pierce, N. F., Sack, R. B., Mitra, R. C., Banwell, J. G., Brigham, K. L., Fedson, D. S., Mondal, A. Ann. intern. Med. 1969, 70, 1173. 2. Nalin, D. R., Cash, R. A., Rahman, M. Bull. Wld Hlth Org. 1970, 43, 361. 3. Pierce, N. F., Hirschhorn, N. W. H. O. Chronicle 1977, 31, 87. 4. Hirschhorn, N., Kinzie, J. L., Sachar, D. B., Northrup, R. S., Taylor, J. O., Ahmed, S. F., Phillips, R. A. New Engl. J. Med. 1968, 279, 176. 5. Palmer, D. L., Koster, F. T., Islam, A. F. M. R., Rahman, A. S. M. M., Sack, R. B. ibid. 1977, 297, 1107. 6. Davidson, G. P., Gall, D. G., Petric, M., Butler, D. G., Hamilton, J. R. J. clin. Invest. 1977, 60, 1402. 7. Tallett, S., MacKenzie, C., Middleton, P., Kerzner, B., Hamilton, R. Pedia-

trics, 1977,60,217. 8.

Yolken, R. H., Kim, H. W., Clem, T., Wyatt, R. G., Kalica, A. R., Chanock, R. M. Lancet, 1977, ii, 263. 9. Sack, D. A., Sack, R. B. Infect. Immun. 1975, 11, 334. 10. Morris, G. K., Merson, M. H., Sack, D. A., Wells, J. G., Martin, W. T., DeWitt, W. E., Feeley, J. C., Sack, R. B., Bessudo, D. M. J. clin. Micro-

biol. 1976,3,486. 11. Nelson, N. J. biol. Chem. 1944, 153, 375. 12. Bishop, R. F., Davidson, G. P., Holmes, I.

H., Ruck, B. J. Lancet, 1973, ii,

1281. 13. Flewett, T. H., Bryden, A. S., Davies, H. ibid. 1973, ii, 1497. 14. Kapikian, A. Z., Kim, H. W., Wyatt, R. G., Cline, W. L., Arrobio, J. O., Brandt, C. D., Rodriguez, W. J., Sack, D. A., Chanock, R. M., Parrott,

Methods Six patients, aged 13-65 years, were studied (see table). Although all were on more than 120 units of insulin daily (given subcutaneously in the thigh or in the arm), their bloodglucose concentrations were high. None of the patients had lipodystrophy. All had been attending the diabetic clinic at the Royal Free Hospital at three-monthly intervals for 2 years or more; the quality of diabetic control had been poor throughout this period. An attempt had been made to improve the diabetic control of these patients by careful regulation of diet and insulin dose before admission. After admission to hospital, the insulin dose was maintained at pre-admission levels for 48 h. Insulin infusion was started on the third day at 8 A.M., before breakfast. The insulin regimen was predetermined and consisted of a basic total dose of 50 units/day, infused at varying rates during the 24 h (fig. 1), the rate being increased at times of the main meals. Overweight patients, who were seen to be qnnrle

Cnnr4

"nn..."

R.H. New Engl.J. Med. 1976, 294, 965. 15. Ryder, R. W., Sack, D. A., Kapikian, A. Z., McLaughlin, J. C., Chakraborty, J., Rahman, A. S. M. M., Merson, M. H., Wells, J. G. Lancet, 1976, i, 659. 16. Echevarria, P., Ho, M. T., Blacklow, N. R., Quinnan, G., Portnoy, B., Olson, J. G., Conklin, R., Dupont, H. L., Cross, J. H. J. infect. Dis. 1977, 136, 383. 17. Shepherd, R. W., Truslow, S., Walker-Smith, J. A., Bind, R., Cutting, W., Darnell, R., Barker, C. M. Lancet, 1975, ii, 1082. 18. Stuart, H. S., Stevenson, S. S. in Textbook of Pediatrics (edited by W. E. Nelson); p.40, Philadelphia, 1959.

LOW-DOSE INSULIN INFUSIONS IN DIABETIC PATIENTS WITH HIGH INSULIN I-Insulin infusion in 24 h).

Fig.

REQUIREMENTS P. DANDONA F. HEALEY

M. FOSTER E. GREENBURY A. G. BECKETT

Royal Free Hospital, Pond Street,

London NW3

Extra insulin was was 63 units in 24 h.

regimen for 60 kg person (total 50 units

given to overweight patients. Maximum

dose used

2QG

patients with high insulin requirements (range 120-3000 units daily) have been infused with much smaller doses (range 50-63 units daily) of insulin intravenously. All six maintained adequate glucose homœostasis on this regimen. It is suggested that subcutaneous tissue at the site of injection may alter insulin or impair its absorption. Insulin resistance in some patients may be due to these mechanSummary

Six

isms. Introduction

INSULIN resistance, defined as a daily requirement of insulin in excess of 200 units, is uncommon and is regarded as being frequently due to’insulin antibodies; it often responds to steroids. However, we had a patient

Fig. 2-Serial blood-glucose concentrations during insulin infusion over 24 h in one patient. Plasma growth hormone and cortisol levels were also measured in this patient.

284

*

Blood-samples taken in the clinic at 9.30-11A.M.

after insulin and breakfast; results show range over 12 mo before mfusion study.

eating larger meals, were given an extra 0.5units/h during the day and an extra 1.5units/h during the 2 h around the meal times. The insulin was given by an intravenous infusion set and in normal saline with 10 ml of plasma protein fraction per 500 ml of saline. The infusion-rate was controlled by an Abbott counting chamber. All patients who had been previously prescribed a fixed carbohydrate diet were allowed a free caloric and carbohydrate intake similar to that taken at home even if it did not correspond with what had originally been prescribed. In each study, lasting 24-48 h, eight plasma-glucose estimates were made between 7.30 A.M. and 11 P.M.; these included three 1 h after breakfast, lunch, and supper, respectively (fig. 2). Urine-glucose was measured before each blood-sample was taken. At the end of the infusion, the patients were put back on doses of insulin which were 20% lower than the original ones. Thereafter the insulin dose was regulated according to the biochemical results. The patients were observed for at least 1 week after the infusion in order to compare the quality of the diabetic control and the requirement of insulin during the infusion with those after the resumption of subcutaneous injections of insulin.

Results

During the insulin infusion the plasma-glucose of all patients was significantly improved. The high glucose concentrations (greater than 12.5mmol/l, fasting) gradually returned to normal within 3-4 h of the start of the infusion and remained reasonable thereafter, the glucose concentrations ranging from 3.5 to 10 mmol/1 (fig. 3). Post-prandial glycosuria did not occur in three cases. Traces of glucose were observed in the urines of the other three but their 24-h urine glucose was less than 800 mg, although two out of the three patients had glycosuria >8 g/24 h when they were on subcutaneous in-

jections of insulin. Ketonuria was not seen in any patient. No hypoglycaemia, biochemical or clinical, occurred. The quality of diabetic control before admission to hospital did not differ significantly from that in the first 48 h of hospital admission (before insulin infusion). After the resumption of subcutaneous insulin at the end of the infusion some patients needed a slightly smaller dose of insulin than previously (5-10% less) and the diabetes remained better controlled than before the infusion. However, within a week of the resumption of subcutaneous insulin injections the insulin requirements rose to pre-infusion levels and diabetic control became erratic. Discussion

patients whose diabetes was poorly controlled despite large doses of subcutaneous insulin, small amounts of intravenous insulin produced adequate glu-’ cose homceostasis, far superior to that produced by much larger doses of subcutaneous insulin. The quality of control was consistent in all the patients studied and was not affected by age, sex, body-weight, duration of diabetes, or original dose of insulin. It may be deduced that the high insulin requirement of the patients was not due to any form of end-organ resistance nor to any significant inhibition of insulin by circulating antibodies or antagonists. Very low titres of insulin antibodies were detected in the three patients in whom they could be measured (insulin binding capacity of 12, 25, and 40 fLu/ml respectively). In the three others, they could not be measured. Hence it would appear that insulin was only partially absorbed from subcutaneous tissue and that the degree of absorption In these

from the subcutaneous site determines exogenous insulin requirements. Subcutaneously injected insulin which is not absorbed is probably degraded locally. Although insulin absorption from the subcutaneous tissues is known to be often erratic3 there have been no observations which suggest that the absorption may not be complete. Insulin degradation has been shown in many organs including liver, kidney, muscle, and brain.’ Little, however, is known about the possible role of subcutaneous tissue on insulin breakdown. It, is possible that in patients such as ours the subcutaneous tissues may act as a mechanical barrier and as an active site of

degradation. Clearly, therefore, Fig. 3-Blood-glucose profiles during insulin infusion. All infusions started at time 0.

insulin resistance can be divided into at least two different types: that where the defect is in the absorption from the subcutaneous tissues, and that where there is resistance to insulin infused intravenously. Patients with high concentrations of insulin

285 those with antibodies to insulin receptors,5 and those with insulin resistance due to subnormal receptors6 probably belong to the second category. It would nevertheless be of considerable interest to study the response of these patients to insulin infusion. Fur-

antibodies,’

thermore,

a

similar

study

in non-insulin-resistant pa-

tients who have usual insulin requirements and in patients with brittle diabetes would clarify whether there is a qualitative difference between them and those with extremely high insulin requirements. Since the control of the diabetes in our patients was more than adequate, it is suggested that this type of insulin infusion might be tried for longer in mobile patients, not only those with insulin resistance but also those with brittle diabetes and indeed in all patients with insulin-dependent diabetes. Albisser et al. used the artificial pancreas to treat brittle diabetes with doses of 27-87 units insulin infused over 8 h, doses which were much larger than those we used. The control of the diabetes improved in two of their three patients. At one stage the third had a bloodglucose of 45 mg/dl (2.5 mmol/1) despite the ability of the artificial pancreas to deliver variable amounts of glucose when necessary; none of our patients, however, had

hypoglycsemia. Carefully regulated insulin infusion may therefore provide a simpler, more practical, albeit possibly not so precise, alternative to the artificial pancreas in the long-term treatment of ambulant diabetic patients. We thank Dr

J. D.

N. Nabarro for his comments and Mrs A. Wins-

bury and Mrs M. Elleman for secretarial help. REFERENCES 1. Oakley, W. G., Jones, V. E., Cunliffe, A. C. Br. med. J. 1967, ii, 134. 2. Page, M. M., Alberti, K. G. M. M., Greenwood, R. et al. ibid. 1974, ii, 687. 3. Binder, C. Absorption of Injected Insulin; p. 50. Copenhagen, 1969. 4. Kitaochi, A. E., Steutz, F. B. Diabetes, 1972, 21, 1091. 5. Flier, J. S., Kahn, C. R., Roth, J., et al., Science, 1975, 190, 63. 6. Oseid, S., Beck Nielsen, H., Pedersen, O., et al. New Engl. J. Med. 1977,

296, 245. 7.

Albisser, A. M., Leibel, B. S., Ewart,

T.

G.,

et

al. Diabetes, 1974,

23, 397.

cal methotrexate and/or cranial irradiation, mean activity in these 16 patients rose significantly to 41±11%. More than 2 years after treatment P.C.A. decreased towards normal in the older children but remained high in the younger group. 5 children with neurological sequelæ (including 3 with A.L.L. and the post-irradiation syndrome) had the highest activities. Ether extraction showed that the active material had lipid properties. P.C.A. did not correlate with protein, lactic dehydrogenase, or cell count in C.S.F.

Introduction BRAIN tissue contains large amounts of thromboplastin1 which may be released during various diseases. Cerebrospinal fluid (c.s.F.) is readily accessible for testing and may reveal changes in the brain. We have examined the procoagulant activity (P.C.A.) of c.s.F. to determine if it is a sensitive indicator of centralnervous-system (c.N.s.) damage.

Patients Normal subjects were divided into groups IA (19 neonates 1-7 days), IB (9 infants aged 1-6 months), and ic (22 children aged 6 months to 17 years). Children with C.N.S. infection were divided into groups IIA (13 children aged 6 months to 15 years studied before therapy) and us (8 children from group IIA studied after therapy for C.N.S. infection). Children with acute lymphatic leukaemia (A.L.L.) were divided into groups IIIA (16 children aged 6 months to 17 years with A.L.L. at diagnosis without c.N.s. involvement), IIIB (group IIIA studied during c.N.s. prophylactic therapy with cranial irradiation and/or intrathecal methotrexate), and IHC (34 children aged 6 months to 17 years with A.L.L. studied more than 2 years after C.N.S. prophylactic therapy identical to that given

aged

to IIIB).

Lumbar punctures were done on the newborns in group IA rule out sepsis; all had Apgar scores >6 at 5 min of life and all had normal c.s.F. findings for age. In the normal infants and children in groups IB and ic, diagnostic lumbar punctures were done to rule out C.N.S. disease. Their symptoms and diagnoses included fever of unknown origin, cervical muscle spasm, headaches, and breath-holding spells. All had completely normal c.s.F. findings and no clinical or laboratory evidence of c.N.s. disease was found in any patient in this group on fol-

to

low-up. SPINAL-FLUID PROCOAGULANT ACTIVITY: A SENSITIVE INDICATOR OF CENTRAL-NERVOUS-SYSTEM DAMAGE

JANET E. GRAEBER

(1/13). Methods

MARIE

J. STUART Department of Pediatrics, State University of New York, Upstate Medical Center, Syracuse, New York, U.S.A.

Testing the procoagulant activity (P.C.A.) of cerebrospinal fluid (C.S.F.) by measuring its effect on plasma-recalcification time is a useful indicator of central-nervous-system damage. C.S.F. from 22 normal children and adolescents aged 6 months to 17 years had a mean ±S.D. P.C.A. of 14±6%. P.C.A. in 13 children with C.N.S. infection was significantly increased to 59±13%. In 8/13 of these children activity remained high (42±11%) after therapy. In patients with acute lymphatic leukæmia (A.L.L.) 12 aged >2½ years at diagnosis had a normal activity (17±8%) and 4 patients aged