HOW LIQUORICE WORKS

1206 SICK DAYS IN GREEN STREET ELEMENTARY SCHOOL,

1986-87

*Statistical comparisons between the number of sick days are by X’ test with one degree of freedom. tFor FSF light vs rest of school (and vs three paired rooms).

In the term before the FSF light was introduced the sickness rate in the FSF designated classrooms was not significantly different from that in the rest of the school or in three paired classrooms. Thus there was no evidence that students in the experimental rooms started off any healthier than the other children. When the FSF light was in place, the sickness rate in the experimental classrooms was lower than that in the rest of the school and in the three paired classrooms (table). The effect was to reverse the usual seasonal pattern in the school wherein sickness absences increase during the winter and spring months. In classrooms without FSF light the rate of sick days increased significantly from September-December to January-June period but in the FSF light classrooms the sickness rate fell slightly. This study was not blind but neither the pupils nor the staff expected that FSF light might affect sickness absence, and nor did I. DSF light raises the serum cortisol level1 more than FSF light doesand glucocorticoids suppress cell-mediated immunity.3 Perhaps FSF light, which is used to treat the seasonal affective disorder,4 may also be useful in the treatment of immune disorders. The teachers found the FSF light brighter, more natural, and more pleasant, and they tended to keep it on for longer and did not wish to return to the DSF light. I thank Mr Robert Neubauer, principal, and Mrs Marie Tatro, secretary, of the Green Street Elementary School. 6 Tyler Street, Brattleboro, Vermont 05301, USA

WAYNE P. LONDON

1. Hollwich F, Dieckhues B, Schrameyer B. Die Wirkung des naturlichen und kunstlichen lichtes uber das auge auf den hormon-und stoffwechselhaushalt des menschen. Klin Mbl Augenheilk 1977; 171: 98-104. 2. Hansen T, Bratlid T, Lingjarde O, Brenn T. Midwinter insomnia in the subarctic region: evening levels of serum melatonin and cortisol before and after treatment with bright artificial light. Acta Psychiatr Scand 1987; 75: 428-34. 3. Calabrese JR, Kling MA, Gold PW. Alterations in immunocompetence during stress, bereavement, and depression: focus on neuroendocrine regulation. Am J Psychiatry 1987; 144: 1123-34. 4. Lewy AJ, Sack RL. Light therapy and psychiatry. Proc Soc Exp Biol Med 1986; 183: 11-18.

(SD) whole-blood aluminium in preterm infants. 0 = group A, fed parenterally; . = group B, fed standard cows’ milk formulae. Figure in parenthesis = number of samples. Significant difference between the groups at weeks 1-2 (p < 0-01), 3-4, and 5--6 (p < 0-05, Wilcoxon rank-sum test). Mean

(range over weeks 3-6 was 9-65 µg/1). In adults with advanced renal disease blood levels of over 50 ug/1 have been associated with osteodystrophy and encephalopathy.4 The effects of serum aluminium concentrations approaching this level in preterm infants with a high bone turnover and a developing nervous system are unknown. Brain aluminium concentrations are increased in neonatal uraemia.5 Our two groups were not strictly comparable: the parenterally fed babies were more preterm (mean 28-1vs 30-5 weeks) and lighter at birth (1136 g vs 1465 g) than those fed enterally. The two groups were not fed similar quantities of aluminium: group A received a mean of 17-6 ug/kg per day (range 0-8-38-8) whereas group B received 47-8 µg/kg per day (17-0-59-2). Consequently we do not know whether the blood levels reflect changes in aluminium absorption or excretion, and longitudinal studies are in progress. Neonatal Intensive Care Unit,

Hope Hospital, Salford M6 8HD

BLOOD ALUMINIUM LEVELS IN PRETERM INFANTS FED PARENTERALLY OR WITH COWS’ MILK FORMULAE

SIR,—We1 and others2 have reported widely variable aluminium levels in milk and parenterally administered solutions. Preterm infants, who have poorer renal function, are at risk of the toxic effects of excess aluminium retention, including encephalopathy.3 Between April and November, 1986, 56 blood samples were collected from 22 preterm infants (24-32 weeks’ gestation). Blood was withdrawn into aluminium-free bottles via stainless steel needles, and samples were analysed in a single batch by graphitefurnace, atomic absorption spectroscopy. The infants were divided into two groups: group A was fed parenterally and group B was fed standard cows’ milk formulae, although some in group B initially received glucose and electrolyte infusion. Mean whole-blood aluminium levels are shown in the figure. At birth mean aluminium levels were similar in the two groups, as expected (about 18 ug/1). There were significant differences in whole-blood aluminium levels at weeks’ 1-2, 3-4, and 5-6. No patient received parenteral nutrition for more than 6 weeks. Blood aluminium levels of bottle-fed babies fell over the 6-week period, approaching values previously obtained in term infants (2-0-10-0 ug/1).3 Mean blood levels in parenterally fed infants increased over the first 2 weeks to reach a plateau at about 28 pg/l

M. J. ROBINSON S. W. RYAN C. J. NEWTON

J. P. DAY Department of Chemistry, University of Manchester

C. D. HEWITT M. O’HARA

1. McGraw M, Bishop N, Jameson R, et al. Aluminium content of milk formulae and intravenous fluids used in infants. Lancet 1986; i: 157. 2. Weintraub R, Hams G, Meerkin M, Rosenberg AR. High aluminium content of infant milk formulas. Arch Dis Child 1986; 61: 914-16. 3. Sedman AB, Klein GL, Meritt JR, et al. Evidence of aluminium loading in infants receiving intravenous therapy. N Engl J Med 1985; 312: 1337-43. 4. Parkinson IS, Ward MK, Feest TG, Fawcett RWP, Kerr DNS. Fracturing dialysis osteodystrophy and dialysis encephalopathy: an epidemiological survey. Lancet 1979; i: 406-09. 5. Freundlich M, Zilleruelo G, Abitbol C, Strauss J, Faugere M-C, Malluche HH. Infant formula as a cause of aluminium toxicity in neonatal uraemia. Lancet 1985; ii: 527-29.

HOW LIQUORICE WORKS

SIR,—Dr Stewart and his colleagues (Oct 10, p 821) postulated that liquorice acts by inhibiting 11 &bgr;-hydroxysteroid dehydrogenase (11&bgr;-OHSD), a microsomal enzyme complex which catalyses the reversible conversion of mineralocorticoid active cortisol to mineralocorticoid inactive cortisone. It is suggested that raised intrarenal cortisol levels saturate renal extravascular cortisolbinding globulin, leaving cortisol to act on the mineralocorticoid receptor. If so one would expect anyone consuming liquorice to be

1207 risk of "apparent mineralocorticoid excess", provided they took enough of it. Stewart and colleagues gave 200 g liquorice daily (containing 580 mg glycyrrhizic acid). Was this amount chosen in an attempt to avoid significant biological effects? It is stated that glycyrrhizic acid and its hydrolysis product glycyrrhetinic acid (the active components of liquorice) in amounts as small as 700 mg daily (as glycyrrhizic acid) cause sodium retention, resulting in hypertension, hypokalaemia, and suppression of plasma renin activity and aldosterone levels.Yet Cumming et aP have reported a patient who presented with flaccid quadriplegia and a hypokalaemia (12 mmol/1) after taking as little as 150 mg glycyrrhizic acid per week. This patient was further investigated under controlled conditions (fixed intake of sodium and potassium and after a run-in period; liquorice containing 94 mg calcium and potassium salts of glycyrrhizic acid per day). During the 11-day period when liquorice was given the urinary potassium loss was slight and less than the loss in the run-in period. However, the patient was in net negative potassium balance because of increased faecal loss. The liquorice did not provoke diarrhoea; serum potassium fell below 3 mmol/1 and total exchangeable potassium fell from 1464 to 1144 mmol. Stewart et al did not measure faecal electrolytes but their figure suggests that the daily urinary potassium loss never exceeded daily intake. Since the serum potassium fell in all subjects from 411 to 3.69 mmol/l, a substantial faecal loss must have taken place. If inhibition of 11&bgr;-OHSD is the only way in which liquorice exerts its effect, the enzyme must be widely distributed in the bowel

at

wall. First University Clinic of Internal Medicine,

Aarhus Kommunehospital, DK-8000 Aarhus C, Denmark

LISELOTTE LONKA ROBERT SMITH PEDERSEN

Epstein MT, Espiner EA, Donald RA, Hughes H Effect of eating liquorice on the renin-angiotensin-aldosterone axis in normal subjects. Br Med J 1977; i: 488-90. 2. Cumming AMM, Boddy K, Brown JJ, et al. Severe hypokalaemia with paralysis induced by small doses of liquorice Postgrad Med J 1980, 56: 526-29. 1.

SIR,-Dr Stewart and his colleagues suggest liquorice-induced deficiency of 11 &bgr;-hydroxysteroid dehydrogenase (11&bgr;-OHSD) as the cause of the mineralocorticoid effects of glycyrrhetinic acid. The evidence presented strongly supports that conclusion. We would like to recall some older data on the action of glycyrrhetinic acid. In 1953 J. Groen and colleagues in Amsterdam (Ned Tijdschr Geneesk 1953; 97: 3290) were the first to demonstrate that the mineralocorticoid action is closely related to the presence of the 11-oxo group in glycyrrhetinic acid. They showed that analogues without this 11-oxo group lacked mineralocorticoid activity in volunteers. Five years later L. M. Atherton from Edinburgh reported glycyrrhetinic acid inhibited the metabolism of progesterone and 11-deoxycorticosterone whereas the 11-deoxy analogues did not (Biochem J 1958; 69: 75). These studies underlined the importance of the 11-oxo group for the biological action of glycyrrhetinic acid. We suggest that 11&bgr;-OHSD deficiency during liquorice ingestion is caused by competition between glycyrrhetinic acid and cortisol (or some other steroids) that need this enzyme for their metabolism. Department of Medicine, University Hospital Nijmegen, 6500 HB Nijmegen, Netherlands

with impaired renal function from any cause may be complicated by hyperkalaemia. Potassium-sparing diuretics should be administered to such patients only after careful consideration and need to be carefully monitored. Renal Unit, Belfast City Hospital, Belfast BT9 7AB

MARY G. MCGEOWN

POTENTIAL ANALGESIC CONTRIBUTION FROM MORPHINE-6-GLUCURONIDE IN CSF

SIR,-Morphine is metabolised by conjugation with glucuronic to give morphine-3-glucuronide (M3G) and/or morphine-6glucuronide (M6G). M6G is analgesic in mice’ and it binds strongly to opioid receptors.2 Clinical use of morphine may be complicated by the presence of this metabolite. Patients with renal failure need lower doses of morphine to achieve analgesia3and these exaggerated opiate effects may be due to accumulation of M6G.4 Substantial plasma concentrations of M6G have been found after intravenous5 and long-term oral 67 morphine, but what has not been established is the contribution M6G makes to the analgesic effect in man. 15 patients undergoing transurethral prostatectomy were given 10 mg morphine sulphate intramuscularly and 11 were given 30 mg morphine orally as a sustained release preparation (’MST Continus’). Paired samples of CSF and plasma, taken 95 min after the intramuscular dose and 140 min after the oral dose, were analysed by radioimmunoassay for morphine (’Coat-a-Count’; Diagnostic Products Corp, Los Angeles) and for M3G and M6G.5 Sampling times were chosen to allow for the absorption of morphine from sustained release tablets. There were no significant differences in age and weight between the two groups. Morphine, M3G, and M6G were detectable in plasma and in CSF after both intramuscular and oral morphine (table i). Plasma concentrations of M3G and M6G were significantly higher after oral administration; conversely, CSF metabolite concentrations

acid

similar for both oral and intramuscular doses but CSF concentrations were significantly higher after intramuscular administration. The concentration of M6G in CSF, as a mean percentage of CSF morphine concentrations, was 12.6 (SE 3-7) % after an intramuscular dose and 13-3 (4-1) % after the oral dose. Both glucuronide metabolites were found in CSF, and the active metabolite (M6G) was found in concentrations of about 1 nmol/l, which is about 13% of CSF morphine concentrations. M6G is not established as having analgesic activity in man. However, it binds quickly to opioid receptors in bovine brain;2 its analgesic potency in rodents exceeds that of morphine after both subcutaneous administration (3-7-fold) and injection into the cerebral ventricles (45-fold)’ and M6G penetrates the blood-brain barrier in rats, without prior hydrolysis.8 If the central-nervous-system potency of M6G in rodents were to be verified in man, our results would suggest that some 85 % of the analgesic activity of morphine derives from M6G (table n).

were

morphine

TABLE I-CSF AND PLASMA CONCENTRATIONS

W. H. L. HOEFNAGELS P. W. C. KLOPPENBORG

SPIRONOLACTONE AND HYPERKALAEMIA IN PATIENTS WITH IMPAIRED RENAL FAILURE

SIR,-Mr O’Reilly and colleagues’ case (Oct 10, p 859) of life-threatening hyperkalaemia after bladder decompression for high pressure chronic retention has a wider implication. Serious hyperkalaemia developed in their patient on the sixth day after insertion of the urethral catheter, despite good urinary output and decreases in serum urea and creatinine concentrations. Renal function remained impaired with a serum creatinine concentration of 405 µmol/1. Spironolactone (which the patient had been taking for several months) was continued until serum potassium reached 7 mmol/1, when the drug was stopped. Thereafter serum potassium became normal. The administration of spironolactone to patients

Values

as mean

(SE) Significance tested by Mann-Whitney U

test.

TABLE II-RELATIVE ANALGESIC CONTRIBUTIONS

*From mean relative percentages from intramuscular and oral doses combined. tShunomura et al.’ relative concentration multiplied by relative potency. Toxal is sum of analgesic activity of morphine plus M6G.