IN SEARCH OF THE SOMOGYI EFFECT

700 BP,

LABORATORY

RESULTS, AND BODY WEIGHT, BEFORE, DURING,

AND AFTER TREATMENT OF DRUG-RESISTANT SEVERE ESSENTIAL

HYPERTENSION WITH FRUSEMIDE

Results as meaniSD. * BPs were lowered significantly during frusemide therand for lying BP and p<0.02 diastolic, p<0.005) apy (systolic, p<0.001; while uric acid p<0.005 standing. Serum potassium levels fell (p
surprising; besides the workers we cited, 1,2 others3,4 have provided good evidence that the hypotensive action of diuretics is limited by the consequent rise in renin and hence in angiotensin n. This was, as we stated, a major theoretical reason for using captopril together with frusemide. We further showed that, despite large doses of frusemide, captopril effectively suppressed plasma angiotensin n in the long term. Neither of our studies was a formal controlled trial. We did not examine the ability of other drugs (such as beta-adrenergic blockers or methyldopa) to lower plasma angiotensin 11 in combination with diuretics although, a priori, captopril seems likely to be more effective. Moreover, we did not add captopril after starting frusemide because of the danger of inducing severe hypotension.5 There is little doubt, however, from our studies, that the combination of frusemide with captopril is markedly and consistently effective in resistant hypertension, in cases where a loop diuretic given with other drugs has been unsuccessful.

µmol/1 creatinine=0.0113 mg/ dl.

(< 160/100 mm.Hg) was achieved or side-effects occurred. The study was evaluated when the last patient had been treated with frusemide for 4 weeks, which gave a mean treatment period of 12 weeks (range 4-27) and a mean frusemide dose of 75 mg/day (range 40-120). At that time frusemide was discontinued in all the patients and measurements were repeated 2 weeks later. During frusemide treatment the BP, particularly the standing pressure, fell strikingly (see table). The serum-potassium fell 11 %, despite potassium supplements in six patients; serum creatinine and uric acid values rose by 27% and 28%, respectively ; serum sodium and albumin and body weight remained unchanged. 2 weeks after frusemide, none of these measurements differed significantly from pre-frusemide values. In mild’ and severe2 hypertension frusemide lowers the blood pressure, and our data suggest that the BP reduction in drug-resistant hypertension can be ascribed to frusemide, which seemingly is what White et al. observed also. The benefit of combining frusemide and captopril seems to be that the metabolic changes induced by frusemide may be prevented and that the drug regimen can be simplified, which is important for patient compliance. If captopril is to be used as drug of first choice in drug-resistant hypertension, it should always be given with a loop diuretic. Medical Department P, Randers City Hospital,

University Hospital of Aarhus, DK-8900 Randers, Denmark

BENT Ø. KRISTENSEN JESPER SKOV

** The Glasgow and Oxford groups’ reply follows.-ED.L SIR,—Dr Kristensen and Dr Skov, in an elegant study, have confirmed the powerful hypotensive action of large doses of frusemide, and they wonder if this might have been the major component of the blood pressure reduction in our patients. 6 of the 11 patients in the Glasgow series and 5 of the 10 in the Oxford series had, before starting captopril, received frusemide in doses at least as big as those used subsequently with captopril, but in combination with other drugs. 1 other Glasgow patient received similar doses of thiazide throughout. In none of these was blood pressure reduction as effective without captopril. This is not 1.

2.

Finnerty FA, Jr, Maxwell MH, Lunn J, Moser M. Long-term effect offrusemide and hydrochlorothiazide in patients with essential hypertension. Angiology 1977; 28: 125-33. Cantarowich F, Benedetti L, Fernandez JC, et al. High dose frusemide and sodium m hypertension. Nephron 1974; 12: 133-39.

M.R.C. Blood Pressure Unit, Western Infirmary, Glasgow G11 6NT

A. B. ATKINSON J. J. BROWN A. F. LEVER J. I. S. ROBERTSON

N. J. WHITE Nuffield Department of Clinical Medicine, John Radcliffe Hospital, Oxford OX3 9DU

B. RAJAGOPALAN H. YAHAYA J. G. G. LEDINGEAM

IN SEARCH OF THE SOMOGYI EFFECT

SIR.—Dr Gale and colleagues (Aug. 9, p. 279) have studied the response to nocturnal hypoglycasmia in two groups of insulintreated diabetics. One group reacted with a rebound hyperglyc2emia (Somogyi) whilst the other remained hypoglyca--mic. Their conclusion-that the endocrine response to hypoglycaemia is not the major cause of the Somogyi phenomenon-is open to debate. We would like to submit alternative interpretations to their findings. (1) Both groups had a mean rise in plasma cortisol, growth hormone and glucagon levels, following the hypoglycaemia, results which are consistent with other reports. 6-8 There is a significant difference in plasma levels of free insulin (expressed in µU/ml rather than mU/ml) between the two groups. Gale et al. attribute the apparent rebound hypoglycaemia to the lack offree insulin. One can equally take the converse view. If the counter-regulatory hormones do indeed have a significant effect, then the group that failed to raise their morning blood sugar can be said to have remained hypoglycaemic because of persistence of higher circulating insulin 1 Leonetti G, Terzoli L, Sala C, Bianchini C, Sernesi L, Zanchetti A Relationship between the hypotensive and renin-stimulating actions of diuretic therapy in hypertensive patients. Clin Sci Mol Med 1978; 55: 307S-09S. 2. Ibsen H, Leth A, Hollnagel H, Kappelgaard AM, Damkjaer Nielsen M, Glese J Reninangiotensin system in mild essential hypertension: The functional significance of angiotensin II in untreated and thiazide-treated hypertensive patients. Clin Sci Mol 3.

Med 1978; 55: 319S-21S. Vaughan ED, Carey RM, Peach MJ, Ackerly JA, Ayres DT. The renin response to diuretic therapy: a limitation of antihypertensive potential Circ Res 1978; 42: 376-81.

MacGregor GA, Markandu ND, Roulston JE, Jones BE. The antihypertensive action of spironolactone: importance of the response of renin-angiotensin system to the loss of sodium and water. In: Addison GM, Asmussen N, Corol P, Kloppenborg PWC, Norman N, Schroder R, Robertson JIS, eds. Aldosterone antagonists in clinical medicine Amsterdam: Excerpta Medica 1978: 466-74. 5. Atkinson AB, Brown JJ, Leckie B, Lever AF, Morton JJ, Fraser R, Robertson JIS Captopril in a hyponatræmic hypertensive: Need for caution in initiating therapy Lan-

4.

cet 1979; i: 557-58. 6. DeFronzo RA, Hendler R, Christensen N. Stimulation of counter-regulatory hormonal responses in diabetic man by a fall in glucose concentration Diabetes 1980; 29: 125-31. 7. Sonksen PH, Tomkins CV, Srivastava MD, Nabarro JDN Growth hormone and cortisol responses to insulin infusion in patients with diabetes mellitus. Lancet 1972, ii: 155-60. 8. Benson JW, Johnson DG, Palmer JP, Werner PL, Ensinck JW. Glucagon and catechloramine secretion during hypoglycæmia in normal and diabetic man.J Clin Endocrinol Metab 1977; 44: 459-64.

701 levels that has suppressed the tendency to hyperglycasmia. This would explain why one patient exhibited the Somogyi effect when the insulin dosage was reduced. (2) Catecholamines rise in response to hypoglycaemia,9,1O and adrenaline infusion causes an elevation in blood sugar. 11 Furthermore, beta-adrenergic blockade has been claimed to have a stabilising effect on blood sugar levels in poorly controlled

diabetics. 12 Although it

can

be

argued (as in your Aug.

9

editorial)

that these observations do not directly determine the extent of involvement of catecholamines in the pathogenesis of the Somogyi effect, their potential importance may have been insufficiently

in insulin absorption rate remains to be explained. Recent studies indicate that the local subcutaneous blood flow may be an important

determinant.44 These findings imply that

not all episodes of rebound hyperdue to excess insulin dosage and that reduction of insulin dosage will not always produce clinical benefit.

glycaemia

of each hormone alone. 15 None of the above explanations is claimed to be necessarily correct, nor are they mutually exclusive. But if, as is argued, the Somogyi effect is due to low plasma levels of free insulin, it is difficult to understand why a reduction in insulin dosage can sometimes abolish or ameliorate this effect.

Division of Diabetes Endocrinology, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania 19104, U.S.A.

KENNETH R. LYEN DAVID FINEGOLD LESTER BAKER

SIR,-Dr Gale and his colleagues found the so-called Somogyi effect

to be caused by insulin deficiency due to deficient insulin delivery after the hypoglycaemic event rather than by an excess of hormones such as cortisol, growth hormone, counter-regulatory 1

and glucagon. This conclusion accords with our findings from a study of six insulin-treated patients in whom we measured insulin disappearance from subcutaneous tissue on the thigh during 12 consecutive days.2Hypoglycaemia followed by hyperglycxmia was seen only in patients who had an increased insulin absorption during the hours preceding hypoglycaemia followed by a decreased insulin absorption. This indicates that the blood glucose rebound was associated with deficient insulin delivery. The fluctuations in insulin absorption rate were unrelated to the dose of insulin, but rather related to the unpredictable and substantial intraindividual variation in insulin absorption rate seen in some subjects.3 The same blood glucose pattern may, however, also be seen in patients with only minor day-to-day variation in insulin absorption rate. These patients seem to be very sensitive to insulin. The variability

AJ, Cryer PE, Santiago JV et al. The role of adrenergic mechanisms in the substrate and hormonal response to insulin-induced hypoglycemia in man. J Clin Invest 1976; 58: 7-15. 10. Christensen NJ, Alberti KGMM, Brandsborg O. Plasma catecholamines and blood substrate concentrations: studies in insulin induced hypoglycæmia and after adrenaline infusion. Bur J Clin Invest 1975; 5: 415-23. 11 Shamoon H, Hendler R, Sherwin S. Altered responsiveness to cortisol, epinephrine, and glucagon in insulin-infused juvenile-onset diabetics. Diabetes 1980; 29: 284-91. 12. Baker L, Barcai A, Kaye R, et al. Beta adrenergic blockade and juvenile diabetes: acute studies and long-term therapeutic trial. J Pediat 1969; 75: 19-29. 13 Moore RA, Smith RF, Asplin CM. Simple test for nocturnal hypoglycæmia in diabetic patients. Lancet 1979; i: 409-10. 14 Kowarski A, Lacerda L, Migeon CJ. Integrated concentration of plasma aldosterone in normal subjects: Correlation with cortisol. J Clin Endocrinol Metab 1975; 40: 205-10. 15 Eigler N, Sacca L, Sherwin RS. Synergistic interactions of physiologic increments of glucagon, epinephrine, and cortisol in the dog. A model for stress-induced hyperglycemia. J Clin Invest 1979; 63: 114-23. 1. Somogyi M. Exacerbation of diabetes by excess insulin action. Am J Med 1959; 26: 169-91. 2 Lauritzen T, Faber OK, Binder C. Variation in I-125 insulin absorption and blood glucose concentration. Diabetologia 1979; 17: 291-95. 3 Binder C. Absorption of injected insulin. Acta Pharmacol Toxcol 1969; 27: suppl 2.

OLE K. FABER CHRISTIAN BINDER TORSTEN LAURITZEN

Hagedorn Research Laboratory, and Steno Memorial Hospital, Gentofte, Denmark

emphasised. (3) Peak hormonal levels may not be the ideal index for separating the two groups: perhaps total hormonal secretion, such as measurement of urinary hormonal excretion 13 or the area under the curve of serial plasma estimations, 14 might be better. (4) Hormonal interactions could have an effect on blood sugar in excess

are

INTRAVENOUS STOPCOCKS AND PORTS

INJECTION

SIR,-We have already pointed out that the design of the ’Venflon’ intravenous cannula (British Viggo; B.O.C.) can cause unnecessary complications for patients.5 The risk of intermittent intravenous bacterial inoculation exists whenever stopcocks and injection ports are incorporated into cannulae, junctions, or extension tubes. High incidences of bacterial contamination in the interstices of three-way taps have already been reported.6.’ An editorial in the British Medical Journal also warned of the risks of septicaemia from monitoring devices.8 Minute air emboli can be seen to enter via side-ports at each injection and, no doubt, these favour the entry of bacteria. Also, accidental omission of the silicone membrane protecting these ports in the manufacturing plant in Sweden could lead to air embolism in a hypovolaemic or tachypnoeic patient. The Viggo intravenous cannula marketed in the U.S.A. (’Vasculon’) differs from that sold in the U.K. (venflon). The vasculon is radio-opaque and the side-port is firmly and permanently occluded by a plastic stopper, so that it cannot be used for injection purposes (see figure). Why do the manufacturers operate to a double standard? 4. Lauritzen T, Binder C, Faber OK. Importance of insulin absorption, subcutaneous blood flow and residual beta-cell function in insulin therapy. Acta Med Scand (in press). 5. Peters JL, and others. Potential hazards of Viggo intravenous cannular. Lan-

cet 1979, ii: 1239. B, Hargiss C, Schoenknecht FD. Stopcock contamination in

6. McCarthur 7.

ICU. Am J Nursing 1975; 75: 96-97. Dryden GE, Brickler J. Stopcock contamination.

141-42. 8. Editorial Monitoring devices and septicæmia.

Anæsthes

Analg 1979;

an

58:

Br Med J 1979; i: 1748.

9. Garber

Viggo

vasculon and venflon cannula’

side-ports.

showing

open and closed