248

ORAL HYPOGLYCAEMIC DRUGS*By J. D. H. SLATER, M.A., M.B., M.R.C.P.

Medical Registrar, Middlesex Hospital, London

In 1955 Franke and Fuchs published clinicalobservations on the potent hypoglycaemic actionof a sulphonamide, I-butyl 3-sulphanilyl urea,called carbutamide.t After extensive clinical trialsin America, Canada and the United Kingdomcarbutamide has been withdrawn from clinical usebecause of the high incidence of serious toxiceffects. About 5 per cent. of 7,193 patientsdeveloped skin rashes, gastro-intestinal disorders,severe leucopenia, liver damage and generalizedsulphonamide sensitivity reactions; eight patientsdied IO to 30 days after starting treatment(Kirtley, 1957).

In the past two to three years another sulphony-lurea, tolbutamide, which lacks the p-amino groupof the sulphanilamide derivatives, has been widelyused without serious ill-effect in the treatment ofmild diabetes mellitus of the adult type. How thedrug acts is still uncertain, but its clinical useful-ness is well established. This account will, there-fore, be concerned chiefly with the mode of actionand clinical applications of tolbutamide, butreference will be made to chlorpropamide, anothersulphonylurea compound, and DBI, a biguanide,both of which have recently become available forclinical trial.

BackgroundSoon after the discovery of insulin by Banting in

I924, the hypoglycaemic effect of synthalin, abiguanide, was reported by Frank, Nothman andWagner (1928) and studied later by Graham andLinder (I928). It was soon abandoned because itcaused severe liver damage in therapeutic doses(Bertram, 1927). Extracts of plants have beenused for a very long time as traditional remediesfor diabetes in many parts of the world. Blueberryleaf extracts (Myrtillin) were investigated by Allenin I927 and periwinkle, mistletoe and the nickerberry (long popular as 'doctor' bush teas inJamaica) by Hugh-Jones in I955. These, andother herbal remedies are either very toxic(especially to the liver) or have no hypoglycaemic

*This article was received for publication in DecemberI958.tAlso known as BZ55, U6987, Invenol and Nadisan.

effect. In large doses oestradiol benzoate mayreduce the hyperglycaemia of post-menopausalwomen with diabetes (Gessler et al., 1939). Injec-tions of dimercaprol sometimes cause a fall of bloodand urine sugar concentration in diabetics re-ceiving large doses of insulin (Butterfield, 1957).Salicylates also lower the blood sugar level inpatients with diabetes but only in amounts whichcause side effects (Reid et al., 1957).

In 1942 Janbon and his colleagues discoveredthe dramatic hypoglycaemic action of a sul-phonamide, p-amino benzene sulphonamido iso-propyl thiodiazole (IPTD), while assessing itsantibacterial properties in typhoid fever. A few ofthe patients died from hypoglycaemia. BetweenI942 and I946 studies on the mode of action ofIPTD and other thiodiazole compounds in dogsand rabbits were made by Loubatiere at Mont-pellier. He was able to show that IPTD (I) cannotproduce hypoglycaemia after removal of the wholepancreas, but will do so if as little as one-tenth ofthe gland remains, (2) is most effective wheninjected directly into the pancreatic artery orthrough the duct of Wirsung under pressure,(3) acts independently of the nervous system,(4) lowers the blood sugar in roughly inverseproportion to the blood sulphonamide level, and(5) raises the respiratory quotient after glucoseadministration. He postulated a pancreatotropicmechanism to explain these effects.

Pharmacology of TolbutamideTolbutamide (Rastinon, Orinase, D.86o, U2043)

is i-butyl 3p-tolyl-sulphonylurea. It has thefollowing formula:

CH3\ S SO,.NH.CO.NH.C4H,'Unlike sulphanilamide derivatives it has nop-amino group and, therefore, does not give theBratton Marshall reaction but it can be measuredspectrophotometrically in alcohol or chloroformextracts. It has no antibacterial activity and littleor no effect on thyroid function (McGavack, I957).Absorption from the gut is rapid and maximalblood levels are reached in about three hours. The

Protected by copyright.

on Decem

ber 29, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.35.403.248 on 1 May 1959. D

ownloaded from

P

rotected by copyright. on D

ecember 29, 2020 by guest.

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.35.403.248 on 1 M

ay 1959. Dow

nloaded from

Protected by copyright.

on Decem

ber 29, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.35.403.248 on 1 May 1959. D

ownloaded from

May I959 SLAT.ER: Oral Hypoglycaemic Drugs- 49

biological half-life is four to eight hours with muchindividual variation (Baird and Duncan, I957) andthe drug is eliminated by the kidneys as the freelysoluble carboxylic ester. This metabolite has nohypoglycaemic action but is responsible for io to28 per cent. of the total blood tolbutamide level(Stowers, Mahler and Hunter, 1958). There is nodanger of crystalluria. A white precipitate maydevelop on testing the urine for protein withsulphosalicylic acid when large amounts ofcarboxy-tolbutamide are being excreted.

Tolbutamide is supplied as i g. white tablets.Tlhe effective dose is I to 3 g. per day, and increas-ing the dose above 3 g. per day will not usuallyproduce a greater fall of blood sugar concentra-tion. Because it is rapidly excreted it should begiven in divided doses every six to eight hours.

Mode of ActionSulphonylurea compounds are now widely used

in selected cases of diabetes mellitus. Alreadyover 300,000 patients all over the world have beentreated with tolbutamide. It is, therefore, mostimportant that the mode of action is clarified andthe relevant clinical and experimental data will beconsidered in some detail. Most of the earlierobservations were made with carbutamide, butsince there is no evidence to suggest that thesulphonylurea compounds have different modesof action, the data obtained with carbutamide willbe included.

It is generally agreed that the sulphonylureacompounds will only produce hypoglycaemia in thepresence offunctioning pancreatic ( cells. Houssayand Penhos (1956) have confirmed that thepresence of some pancreatic tissue is essential forthe hypoglycaemic effects of these compounds inthe same way that Loubatiere showed that this wastrue of the thiodiazole sulphonamides. Thesulphonylureas will reduce the fasting blood sugarlevel of intact and partially pancreatectomizedanimals or animals made mildly diabetic withalloxan. They will not do so after total pan-createctomy, either in animals or man, and theyhave no effect in animals made severely diabeticwith alloxan. In humans the drugs are onlyeffective in individuals who are able to produceinsulin. They will reduce the fasting blood sugarconcentration in non-diabetics and mild diabeticsof the adult type but with very rare exceptions, thedrugs will not act in patients with severe diabetesof the juvenile type. The pancreas of adultdiabetics contains over 30 per cent. of the normalinsulin content, while, in juvenile diabetics, littleor no insulin can be extracted from the gland(Wrenshall and Best, 1956). Similarly, insulin-like activity is readily detected in the plasma ofadult cdabetics but gerne:rally cannot be found in

0v-- +O,0--M + VACTIVE O INACTIVE

T-- + . r--''+INACTIVE ACTIVE

O Inactivating ionV Calcium ion

Active "local factor (or factors)II Inactive "local factor'(or factors)

the plasma of patients with the juvenile form of thedisease (Bornstein and Lawrence, I95I).

After evisceration in dogs (Fritz et al., 1956;Houssay et al., 1957), rabbits (Wick, Britton andGrabowski, 1956) and rats (Lang and Sherry,1956) the drugs will not cause hypoglycaemia.Like insulin, removal of the pituitary or adrenalglands increases the hypoglycaemic effect, but,unlike insulin, hypophysectomized dogs are lesssensitive than adrenalectomized animals (Houssayet al., 1956). In man, Fajans and his colleagueshave shown that the drugs are active in patientswith panhypopituitarism and Addison's diseaseboth before and after replacement therapy withhydrocortisone. The hyperglycaemic response toadrenalin is unaffected and the rise of blood sugarlevel following injected glucagon is not inhibitedin man (Fajans et al., 1956) or rabbits (Berthetet al., 1956).From these results the following conclusions can

be drawn concerning the mechanism by which thesulphonylurea compounds lower the blood sugarlevel: (i) they do not have a direct insulin-likeaction on peripheral tissues,* (2) the presence of aninsulin-containing pancreas is essential, (3) theydo not act by reduction of glucagon secretion or bytissue antagonism of glucagon, and (4) inhibitionof hormonal insulin antagonists from the pituitaryor adrenals is not the mechanism.We are left with three plausible ways of explain-

ing the hypoglycaemic effect of the compounds:(a) the pancreas is stimulated to produce more

*Lundbaeck, Nielsen and Rafaelsen (1958) claim tohave shown that in vitro tolbutamide slightly increasesglucose uptake by the rat diaphragm. This is uncon-firmed and contradicts earlier work using similarm:ethods (Recant and Fischer, 1957; Carhil]. et al., 1957).

Protected by copyright.

on Decem

ber 29, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.35.403.248 on 1 May 1959. D

ownloaded from

POSTGRADUATE MEDICAL JOURNAL

insulin, (b) the rate of endogenous insulin destruc-tion is diminished, and (c) hepatic glucose releaseand/or production is depressed.

(a) Liberation of insulin from the pancreas wasfirst suggested by Loubatiere to explain the hypo-glycaemic effects of IPTD and other thiodiazolesulphonamides. In cross circulation experimentshe was able to show that when IPTD was injectedinto a normal dog whose pancreaticoduodenal veinwas anastomosed to the jugular vein of a de-pancreatized dog, the blood sugar level of thediabetic animal was depressed. Pozza and his co-workers, in similar experiments with intactanimals, have demonstrated that, following intra-venous carbutamide, the blood sugar level of therecipient animal could be lowered by the donor'spancreatic vein blood, but not by blood from thedonor's mesenteric vein. Direct perfusion of thepancreas with tolbutamide via the right gastro-epiploic branch of the gastroduodenal artery hasbeen performed by Colwell, Colwell and Colwell.They have shown a greater fall of blood sugarlevel by this route than by injection into thefemoral vein, but they were unable to producehypoglycaemia by pancreatic injection in doseswhich were too small to be effective when givenperipherally.The response to tolbutamide appears to be pro-

portional to the insulin content of the pancreas.Mirsky, Perisutti and Gitelson have shown in dogsthat pre-treatment with growth hormone, whichis known to deplete the pancreas of insulin(Campbell et al., 1953), will reduce the acutehypoglycaemic effect of tolbutamide. Similarly,prolonged fasting, which also reduces the insulincontent of the pancreas (Best, Haist and Ridout,1939), will diminish the hypoglycaemic response.Dulin and Miller have shown that after hypo-physectomy in rats the amount of insulin in thepancreas falls to about half the pre-operative levelat four weeks, and they claim that sensitivity totolbutamide diminishes correspondingly. Inhumans, the rate of fall of blood sugar concentra-tion following a single dose of tolbutamide (I g.intravenously) can be used as a diagnostic test formild diabetes (Unger and Madison, 1958). Innormal people the fall of blood sugar level is morerapid and more profound than in patients withmild diabetes.

Morphological studies of the pancreatic islettissue in animals show degranulation of the p cellswith swelling of their nuclei after the administra-tion of large amounts of tolbutamide or carbuta-mide; the islets increase in size and number andmitoses are more frequent (Ashworth and Haist,1956; Pfeiffer et al., 1957; Von Holt et al., I956).So far no definite morphological changes in theislet tissue have been observed in the few diabetics

who have died while receiving tolbutamide(Creutzfeldt, 1956; Ferner and Runge, 1956).Pfeiffer and his colleagues have tried to correlatehistological changes in the 1 cells with changes ofblood sugar level and pancreatic insulin content.In experiments with calves they have been able toshow a transient degranulation of the P cells withincrease in nuclear volume and a transient reduc-tion of extractable insulin from the pancreasfollowing a single moderate-sized dose of tol-butamide by mouth. The fall in blood sugar con-centration closely paralleled these changes. Thatp cell degranulation is a specific response to thehypoglycaemic sulphonylurea compounds is sug-gested by Creutzfeldt, Detering and Welte (I957).They have shown that large doses of two othersulphonylureas (N-(4-methyl-benzene-sulphonyl)-N'-methylcarbamide and N-sulphanilyl-N'-ethyl-carbamide) have no effect on the blood sugar leveland do not produce changes in the p cells.

Perhaps the most compelling evidence that thedrugs produce hypoglycaemia by the release ofendogenous insulin is provided by Pfeiffer, Renold,Thorn and others who have confirmed the earlierwork of Recant and Fischer. They have shown,by assay on the rat epididymal fat pad (Martin,Renold and Dagenais, I958), that the insulin-likeactivity of pancreatic vein blood (which is normallyhigher than peripheral vein blood) increases manytimes after the administration of hypoglycaemicsulphonylureas to normal dogs. In rats, one groupof workers find that the insulin-like activity ofperipheral vein blood is increased by carbutamide(Von Holt et al., I957) and tolbutamide (Krachtet al., 1957). But in humans tolbutamide does notproduce a rise of insulin-like activity in the peri-pheral blood when assayed by the glucose uptakeof the rat diaphragm (Renold et al., 1957; Weaveret al., I958).*

If, then, the sulphonylurea compounds producehypoglycaemia by stimulating insulin release fromthe pancreas, it should be possible to demonstratesome of the metabolic effects of increased peri-pheral glucose utilization in man. After insulininjections, the peripheral arterio-venous glucosedifference increases, plasma potassium and phos-phate concentrations fall, lactate and pyruvatelevels rise and there is an increase of the respiratoryquotient. With rare exceptions these changes havenot been observed following tolbutamide or car-butamide, although many observations by different

*Using a more sensitive rat diaphragm techniqueFraser, Joplin and Vallance-Owen have now been ableto detect in the peripheral blood of tolbutamide-sensitive diabetics a significant rise of insulin-likeactivity following a single oral dose. They find a goodcorrelation (r=o.69 (p o.ox) ) between the increase ofglucose uptake by the diaphragm and the fall of bloodsugar concentration.

May I959250P

rotected by copyright. on D

ecember 29, 2020 by guest.

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.35.403.248 on 1 M

ay 1959. Dow

nloaded from

SLATER: Oral Hypoglycaemic Drugs

workers have been made (Purnell et al., 1956;Renold et al., 1957, and others). However, suchnegative findings lose much of their importanceonce it is realized that these indices of peripheralglucose utilization are relatively insensitive andliable to considerable experimental error. Wheninsulin is given intravenously in doses whichmimic the fall of blood sugar level produced bytolbutamide, evidence of increased peripheralutilization is also often inconclusive, especially ifthe injection is made slowly. Madison and Ungerhave shown that when insulin is injected into theportal vein (thus simulating endogenous secretionof insulin) little effect on arterio-venous glucosedifferences can be demonstrated, although equiva-lent doses of insulin by rapid injection into aperipheral vein are followed by an easily demon-strable increase. Because of these difficulties itis perhaps significant that Goetz et al. (1956) doclaim to have shown increased glucose uptake bythe tissues following intravenous tolbutamide inhigh doses. By simultaneous measurement offore-arm blood flow, Butterfield and his co-workers claim to have found that there is a criticalblood sugar concentration below which glucosedoes not enter cells. This threshold is raised indiabetes and can be lowered by oral tolbutamide ifthe drug is therapeutically effective.

In intact animals the sulphonylurea compoundsincrease liver glycogen content without affectingmuscle glycogen; single injections of insulin dothe reverse. However, if a constant intravenousinsulin infusion is given to fasting, intact rats(Dulin and Johnston, I957), the blood sugar falls,although muscle glycogen remains unchanged.This again emphasizes that the metabolic responseto insulin is greatly effected by the route and rateof administration.

(b) Diminished destruction of insulin has beensuggested by Mirsky, who has demonstrated in-hibition of insulinase in rat liver slices. Vaughanhas shown, however, that insulinase is not in-hibited by the blood levels reached in man follow-ing therapeutic doses of tolbutamide. In intactrabbits Weaver and his co-workers have shown thatthe rate of degradation of I 31-labelled insulin(separated from the plasma electrophorectically) isnot altered by tolbutamide, and the liver of ratspre-treated with the drug destroy I131-labelledinsulin at the normal rate (Williams and Tucker,1956). The majority of workers have been unableto obtain potentiation of exogenous insulin by thesulphonylureas in severe diabetics or followingpancreatectomy in man. However, in de-pancreatized dogs there is some evidence that largedoses of carbutamide or tolbutamide (Houssay etal., 1956; Caren and Corbo, I957) will allow areduction of the insulin dose. Diminished des-

truction of insulin, therefore, does not appear tobe an important effect of the sulphonylureas, butas Mirsky suggests, inactivation of insulinase maybe responsible for the slow rise of blood sugarlevel following tolbutamide-induced hypogly-caemia which contrasts with the rapid rise seenafter insulin.

(c) Reduction of hepatic glucose output by a directaction on the liver has been suggested by themajority of workers because the sulphonylureacompounds do not appear to increase glucose up-take by the peripheral tissues in man. There is nodoubt that tolbutamide reduces the output ofglucose by the liver in humans and in dogs, duringfasting and after fructose administration (Renoldet al., 1956; Moorhouse et al., 1957) whethermeasured directly by means of catheterizationstudies or indirectly by the rate of fall of thespecific activity of Ci4-labelled glucose (Ashmore,Cahill and Earle, I957; Tarding and Schambye,1958; Miller et al., 1957; Purnell et al., 1956;Recant and Fischer, I957).

In vitro, tolbutamide interferes with the conver-sion of liver glycogen to glucose. Large amountsadded to the incubating medium prevent glucoserelease from rat and rabbit liver slices (Clark et al.,1956; Vaughan, 1956; Berthet et al., 1956). Theincreased glucose output induced by adrenalin isparticularly affected (Vaughan, 1957). Weber andCantero have shown that tolbutamide inhibitsglucose-6-phosphatase predominantly; phospho-hexosisomerase activity is slightly reduced butliver phosphoglucomutase, glucose-6-phosphatedehydrogenase and 6-phosphogluconate dehydro-genase are unaffected. This suggests that in vitrotolbutamide has a specific effect on glucose-6-phosphatase, although the concentrations neces-sary are five to ten times higher than the peripheralblood levels which produce hypoglycaemia in vivo.Liver tissue removed from rats during tolbutamidehypoglycaemia does not show reduced glucose-6-phosphatase activity (Ashmore, Cahill andHastings, 1956) and it should be remembered thatmany other sulphonamides or sulphonamidederivatives which do not lower the blood sugarwill also inhibit liver glucogenic enzymes.

Dulin and Johnston claim that removal of theliver does not prevent the hypoglycaemic responseto tolbutamide in dogs, but this work has not beenconfirmed.* Since the animals need a constantglucose infusion in order to survive, the interpreta-tion of changes of blood sugar level is difficult.However, supporting evidence is reported byColwell, Colwell and Colwell, who showed that

*Sobel et al. (I958) have now confirmed this finding.Moderate doses of tolbutamide produced a similar fallof blood sugar concentration in liverless and intact dogs.

May I959 251P

rotected by copyright. on D

ecember 29, 2020 by guest.

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.35.403.248 on 1 M

ay 1959. Dow

nloaded from

POSTGRADUATE MEDICAL JOURNAL

intraportal injections of carbutamide did notcause hypoglycaemia and that injections of tol-butamide into the hepatic artery did not produce a

greater fall of blood sugar than peripheral injec-tions. These findings make it difficult to believethat, in therapeutic doses, the sulphonylurea com-pounds reduce hepatic glucose output by a directaction on the liver and it is possible that this effectmay be due largely to increased endogenous insulinsecretion. If not, it is difficult to explain why thesulphonylurea compounds are only effective in thepresence of an insulin-containing pancreas, al-though it is conceivable that the metabolic upsetfollowing pancreatectomy in some way masks aprimary hepatic effect of the drug. However, ourknowledge of the hepatic action of insulin is still in-conclusive. An increase of hepatic glycogen underthe influence of insulin can be demonstratedreadily in intact animals provided the insulin isrigorously glucagon-free and no fall of bloodsugar is allowed to occur. Dunn and his colleagues,using the rate of decay of the specific activity ofCI4-labelled glucose as an index of hepatic glucoseoutput, have observed a' plateau' lasting Io to 20minutes following io units of insulin intravenouslyin unanaesthetized, intact dogs. This indicatessuppression of the entry of glucose into thecirculation. Their work confirms earlier observa-tions made in humans by Beam, Billing andSherlock using catheterization techniques. Morerecently, however, Mahler and his colleagues bycatheter studies in unanaesthetized dogs havefailed to obtain evidence of inhibition of hepaticglucose output following insulin injected into theportal vein or splenic artery. They find that whenthe portal blood sugar concentration falls below60 mg. per Ioo ml., there is an increased output ofglucose from the liver. This does not happen if asimilar fall of blood sugar level has been inducedby tolbutamide. Tarding and Schambye havealso failed to show a reduction of hepatic glucoseoutput following constant intraportal infusions ofsmall amounts of insulin.

In conclusion, although from animal experi-ments the evidence that the sulphonylureas pro-duce hypoglycaemia by releasing insulin from thepancreas is quite convincing, the only evidence ofsuch a mechanism in man is that juvenile diabeticsand pancreatectomized patients do not respond tothe drugs.* Reduction of hepatic glucose produc-tion appears to be the reason for the fall in bloodsugar level after tolbutamide, but whether this iscaused pharmacologically by directly inhibitinghepatic glucogenic enzymes or physiologically byan increased output of endogenous insulin, cannotbe answered until our knowledge of the hepatic*See footnote on page 250.

effect of insulin is more certain. It is possiblethat both mechanisms may operate and a directaction on the liver cells may enhance the hepaticeffect of endogenous insulin.

Clinical ConsiderationsThe sulphonylurea compounds are only, effective

in patients with mild diabetes of the late-onset oradult type; 60 to 70 per cent. of the diabeticpopulation belong to this group. They are notinsulin deficient (as defined above) and they haveno tendency to develop diabetic ketosis. In con-trast, the sulphonylureas are useless in severediabetes of the juvenile type. These patients aretruly insulin deficient and they need daily injec-tions of insulin to maintain health. Giving sul-phonylurea compounds at the same time will notusually allow a reduction of the insulin dose, and alabile or 'brittle' diabetic cannot be made morestable.The majority of patients with mild diabetes of

the adult type are obese, and weight reductionwith adequate restriction of carbohydrate in thediet will often control their symptoms and hyper-glycaemia. There is, however, a small group ofmiddle-aged or elderly diabetic patients, constitut-ing perhaps 5 to 0o per cent. of the total diabeticpopulation whose diabetes has developed late inlife, sometimes with moderately severe symptoms;-they have little or no tendency to ketosis, they areoften under-weight and their hyperglycaemia can-not be adequately controlled by diet alone. Thesepatients would otherwise require insulin and willbenefit most from treatment with the sulphonyl-urea compounds.Selection of Patients

Suitable patients can usually be selected onclinical grounds alone, although this may bedifficult if the patient is already taking insulin.Objective tests have, therefore, been devised toassess the likelihood of a good therapeutic responseto tolbutamide. At the Joslin Clinic (Marble andCamerini-Davalos, 1957) a single 3 g. oral dose oftolbutamide is given when fasting; after fourhours, patients likely to respond to the drug willhave shown a fall of blood sugar (Somogyi-Nelsontechnique) to i o0 mg. per Ioo ml. or less. Patientswith a fasting blood sugar of over 250 mg. pero00 ml. rarely show the requisite fall and are

usually unsuitable for treatment with tolbutamide.This observation is confirmed by experience at theMiddlesex Hospital, where it is found that tolbuta-mide seldom reduces the random blood sugar levelby more than 150 mg. per Ioo ml. Selection bythis method is very stringent 'and reasonablyreliable but some suitable patients will be missed.

May I959252~P

rotected by copyright. on D

ecember 29, 2020 by guest.

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.35.403.248 on 1 M

ay 1959. Dow

nloaded from

SLATER: Oral Hypoglycaemic Drugs

Ducan et a. used the development of ketoacidosisas an index of response to tolbutamide. If ketonesappear in the urine eight hours after the lastinsulin dose, oral therapy is contraindicated, but ifno ketones have appeared within 24 hours, tol-butamide will usually be successful. There are,however, some patients unresponsive to tolbuta-mide who do not easily develop ketosis. Becausethe value of these short-term tests is limited, thebest method of selection is a therapeutic trial. Asa general rule tolbutamide should only be con-sidered for patients with mild diabetes who arewithin io per cent. of the ideal weight for theirheight and years (Kemsley, I95i). Although theymay be receiving insulin, obese diabetics can oftenbe controlled by weight reduction and carbo-hydrate restriction alone once the insulin has beenstopped. If such patients are treated with oralhypoglycaemic drugs, weight reduction becomesalmost impossible. In patients of normal weightwho are taking insulin the insulin dose should begradually reduced until hyperglycaemia appears.Tolbutamide in divided doses of I to 2 g. per daycan then be introduced and the insulin dosereduced still further. If the blood sugar level falls,insulin may be reduced again and finally given upentirely. The patient should be seen daily andfrequent blood sugar estimations performed.Ketonuria must be watched for carefully. Thechangeover is best performed in hospital, par-ticularly in those receiving more than 20 units ofinsulin daily, because some patients rapidly be-come uncontrolled and insulin treatment has tobe resumed at once.A delayed type of response is seen in some

patients. Hyperglycaemia may not respond forseveral weeks although ketosis does not develop.The blood sugar level then falls slowly and there-after random blood sugar levels remain satisfactory(Walker et al., 1957).About 5 to Io per cent. of patients who initially

respond well to tolbutamide cease to do so aftersome months of treatment. Dietary indiscretionswill explain a few of these cases, but there aremany who appear to develop a genuine resistanceto the drug. When assessed in hospital, hyper-glycaemia persists despite large doses and a rigiddiet; and, the drug can be stopped without anyfurther rise of the blood sugar level. As timepasses, increasing numbers of patients who havebecome resistant are being seen. It may representan 'exhaustion' of the pancreatic 3 cells due torepeated stimulation, although Pfeiffer has shownthat the dose of insulin needed afterwards is nohigher than before tolbutamide was begun.Nevertheless, following a period of insulin treat-ment many of these patients will again respondto tolbutamide.

ToxicityAs used therapeutically in man, tolbutamide is

remarkably non-toxic. At the Joslin Clinic sinceNovember I957 only I.I per cent. of 772 patientshave developed side effects sufficiently severe towarrant stopping the drug. Dolger has seen noserious toxic reactions among 500 private patientstreated for 3 to 12 months. Tolbutamide maycause gastro-intestinal symptoms such as epigastricpain, nausea and diarrhoea, and in patients withpeptic ulcer, their dyspepsia may be aggravated.Beaser reports an old man who perforated an acuteduodenal ulcer after three months treatment withlarge doses (2.5 g. per day). Urticarial skin rashesare not uncommon, and a single case of purpurahas been reported by Sugar. Flushing of the facemay occur after alcohol. Most clinical reportsdescribe abdominal symptoms and skin rashes inI to 3 per cent. of patients, but at the MiddlesexHospital 7 per cent. of 20o patients have developedthese features. Most observers, have been unableto detect any significant changes in the white cellcount, but leucopenia has been reported from timeto time. It is usually transient and no case ofagranulocytosis has been recorded. Crosland-Taylor has observed a temporary fall in theneutrophil polymorph count by the third week oftreatment to under 3,000 per c.mm. in 34 per cent.of 32 patients. In two patients leucopenia haspersisted but the neutrophil count has not fallenbelow 1,500 per c.mm. during the I2 monthsof treatment.

Repeated liver function studies have been per-formed by Marble and Camerini-Davalos and bySherry and Zeffren. No changes of the serumbilirubin, flocculation tests and bromsulphaleinexcretion have been seen over a period of I2 to I6months. A slight rise of serum alkaline phospha-tase level has been reported by Marble (1958) butits significance is doubtful. Dolger has giventolbutamide to several patients who had recentlyrecovered from hepatitis or obstructive jaundicewithout evidence of further liver damage. Never-theless, in dogs severe liver cell damage doesdevelop; Sirek and his colleagues have shown thatlarge doses of tolbutamide (30 to Ioo mg. per kg.)given over a period of 14 months produce jaundice,a fall in serum albumin concentration, a rise ofalkaline phosphatase, and a prolonged prothrom-bin time which does not respond to vitamin K.

DangersTolbutamide should only be used when frequent

and detailed observation of the patient is possible.The long-term toxicity is unknown, so that tol-butamide treatment is only justified if hypergly-caemia is adequately controlled. 'A high renalthreshold for glucose makes urine sugar testing a

May I959 253P

rotected by copyright. on D

ecember 29, 2020 by guest.

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.35.403.248 on 1 M

ay 1959. Dow

nloaded from

POSTGRADUATE MEDICAL JOURNAL

poor index of blood sugar concentration in manyof the diabetics who are suitable for tolbutamide,and, therefore, frequent blood sugar estimationsshould be performed. Ketoacidosis may occur atthe time of changing over from insulin, or it mayappear rapidly during the course of an acute infec-tion. Even without ketonuria, serious hyper-glycaemia due to acquired resistance may developinsidiously. Hypoglycaemia is very rare. Obesity,with its attendant dangers, develops easily andexcessive weight gain can only be avoided by care-ful supervision of the patient's diet. That thesulphonylurea compounds do directly inhibitenzyme systems in the liver (and possibly else-where) is important, because any drug whichchronically affects hepatic function would ul-timately be expected to damage the liver cells. Itis too early to say whether the incidence of' degenerative' complications of diabetes mellituswill be affected by treatment with tolbutamide.Better control of hyperglycaemia will presumablytend to reduce the risk of complications but anydrug which increases the tendency to haemorrhagemust be used with great caution, particularly inpatients with retinopathy. Carbutamide increasescapillary fragility (Phemister, 1957), but carefulstudies by Marx have failed to reveal any differencein the incidence of haemostatic abnormalitiesbetween insulin-treated and tolbutamide-treateddiabetes.

Chlorpropamide (P.607, Diabinese)This new and highly potent hypoglycaemic sul-

phonylurea compound was discussed at a meetingof the New York Academy of Sciences in Septem-ber I958. Chemically it is I-propyl-3p-chloro-benzene-sulphonylurea and it has the followingformula:

Cl1 SO.N CONH.CO.NH.C

In contrast to tolbutamide it has a chloride ionin the para position on the benzene ring and apropyl instead of a butyl group as the aryl radical.

Absorption from the gut is rapid and maximalblood levels are reached in two to four hours. Itis not broken down into an inactive metabolite liketolbutamide and is very slowly excreted in theurine. The biological half-life is about 36 hours.This explains why, weight for weight, chlor-propamide has a much more potent hypoglycaemiceffect than tolbutamide. Consequently the doseis much lower and good control can usually beobtained with 200 to 500 mg. daily as a single dose.

Animal studies indicate that chlorpropamideacts in a similar fashion to tolbutamide (e.g. thepresence of the pancreas is essential) and, like the

latter, it is only effective in patients with milddiabetes of the adult type.Whether chlorpropamide has any clinical ad-

vantage over tolbutamide is still undecided. Thedrug is certainly more potent and the blood sugarlevel may be reduced by 200 to 300 mg. per Ioo ml.Generally, however, patients unresponsive to tol-butamide will not do better on chlorpropamide,although diabetics with acquired resistance totolbutamide may respond.

Side effects are quite common with high dosesbut rare with small doses. Prolonged hypogly-caemia is an important danger and, like the hypo-glycaemia of the long-acting insulin preparations,it may only respond slowly to glucose administra-tion. An intrahepatic obstructive jaundice of thetype seen with chlorpromazine may develop evenwith small doses and a curious ataxia with muscleweakness (not due to hypoglycaemia) has beendescribed following large doses of chlorpropamide.Gastro-intestinal side effects are less commoncompared with tolbutamide but an antabuse-likeeffect after drinking alcohol may occur.

Phenethylbiguanide (DBI)The hypoglycaemic effect of this compound was

described by Unger, Freedman and Shapiro inI957. It is not a sulphonylurea and structurallyresembles synthalin. Chemically it has thefollowing formula:

K_ (CH2)2.(NH.CNH)2.NH2The mode of action is quite different from the

sulphonylurea compounds, because the drug iseffective in animals with severe alloxan diabetesand also in some human diabetics of the juveniletype. In vitro it increases glucose uptake by therat diaphragm but the muscle glycogen contentdecreases, there is a marked increase of lacticacid production and oxygen consumption falls(Williams et al., I957). This suggests that DBIstimulates anaerobic glycolysis by inhibitingoxidative enzyme systems. With doses of 150 to200oo mg. per day Krall and his co-workers haveshown that the drug is effective without insulin inmany adult diabetics and occasional children. Noserious toxic effects have been encountered andrepeated liver function studies (over a period ofover x8 months in some cases) have not shown anyabnormality. However, the majority of patientsexperience intolerable anorexia, nausea, vomitingand diarrhoea in doses greater than 200oo mg.per day.This slender margin between the therapeutic

dose and doses which produce severe side effectsmeans that the drug is unlikely to be used exten-sively although Krall et al. have shown that,

May 1959254P

rotected by copyright. on D

ecember 29, 2020 by guest.

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.35.403.248 on 1 M

ay 1959. Dow

nloaded from

May 1959 SLATER: Oral Hypoglycaemic Drugs 255

combined with insulin, DBI will make a labile or'brittle' diabetic more stable. This may proveto be the main clinical application of the drug.

BIBLIOGRAPHYALLEN, F. M. (1927), J. Amer. med. Ass., 89, I577.ASHMORE, J., CAHILL, G. F., and HASTINGS, A. B. (1956),

Metabolism, 5, 774.ASHMORE, J., CAHILL, G. F., and EARLE, A. S. (1957), Ann.

N.Y. Acad. Sci., 71, I3I.ASHWORTH, M. A., and HAIST, R. E. (1956), Canad. med.

Ass. J., 74, 975.BAIRD, J. D., and DUNCAN, L. J. P. (I957), Scot. med. J., 2, 341.BEARN, A. G., BILLING, B. H., and SHERLOCK, S. (1953),

Ciba Foundation Colloquia on Endocrinology, 6, 250.BEASER, S. B. (1957), Ann. N.Y. Acad. Sci., 71, 264.BERTHET, J., SUTHERLAND, E. W., and MAKMAN, M. H.

(1956), Metabolism, 5, 768.BERTRAM, F. (1927), Dtsch. Arch. klin. Med., x58, 76.BEST, C. H., HAIST, R. E., and RIDOUT, J. H. (1939), J.

Physiol. (Lond.), 97, 107.BORNSTEIN, J., and LAWRENCE, R. D. (195I), Brit. med. J.,

i, 732.BUTTERFIELD, W. J. H., and THORNTON, R. H. S. (1957),

Clin. Sci., x6, 679.BUTTERFIELD, W. J. H., KELSEY FRY, I., and HOLLING,

H. E. (1958), Diabetes, 7, 449.CAMPBELL, J., HAUSTER, H. A., MONROE, J. S., and

DAVIDSON, I. W. F. (1953), Endocrinology, 53, 134.CAREN, R., and CORBO, L. (i957), J. din. Invest., 36, 1546.CLARK, D. W., DAVIDSON, M., SCHONBAUM, E., and

SENMAN, H. (1956), Canad. med. Ass. J., 74, 966.COLWELL, A. R., COLWELL, J. A., and COLWELL, A. R.

(1957), Ann. N. Y., Acad. Sci., 71, 125.CREUTZFELDT, W. (1956), Dtsch. med. Wschr., 81, 941.CREUTZFELDT, W., DETERING, L., and WELTE, O.

(I957), Ibid., 82, 1564.CROSLAND-TAYLOR, P., personal communication.DOLGER, H. (1957), Ann. N. Y. Acad. Sci., 71, 275.DULIN, W. E., and JOHNSTON, R. L. (1957), Ibid., 71, I77.DULIN, W. E., and MILLER, W. L. (1958), Proc. i8th Ann. Meet

Amer. Diab. Ass.DUNCAN, G. G., LEE, C. T., and KENT YOUNG, J. (1957),

Ann. N.Y. Acad. Sci., 71, 233.DUNN, D., FRIEDMANN, B., MAASS, A. R., REICHARD,

G., A., and WEINHOUSE, S. (I957), J. biol. Chem., 225, 225FAJANS, S. S., LOUIS, L. H., SELTZER, H. S., GITTLER,

R. D., HENNES, A. R., WAJCHENBERG, B. L., ACKER-MAN, I. P., and CONN, J. W. (1956), Metabolism, 5, 820.

FERNER, H., and RUNGE, W. (1956), Dtsch. med. Wschr., 81, 33r.FRANKE, H., and FUCHS, J. (1955), Ibid., 8o, I449.FRANK, E., NOTHMAN, M., and WAGNER, A. (1926), Ibid.

52, 2067, 2107.FRITZ, I. B., MORTON, J. V., WEINSTEIN, M., and LEVINE,

R. (1956), Metabolism, 5, 744.GESSLER, C. J., HALSTED, J. A., and STETSON, R. P. (x939),

J. din. Intvest., 18, 715.GRAHAM, G., and LINDER, G. C. (I928), Quart. J. Med., 21, 509.GOETZ, F. C., GILBERTSON, A. S., and JOSEPHSON, V

(1956), Metabolism, 5, 788.HUGH-JONES, P. (1955), Lancet, ii, 891.HOUSSAY, B. A., and PENHOS, J. C. (1956), Metabolism, 5, 727.HOUSSAY, B. A., PENHOS, J. C., URGOITI, E., TEODOSIO,

N., APELBAUM, J., and BONKETT, J. (1957), Ann. N.Y.Acad. Sci., 71, 25.

JANBON, M., LAZERGES, P., and METROPOLITANSKI,J. H. (1942), Montpellier Med., 22, 489.

JANBON, M., CHAPTAL, J., VEDEL, A., and SCHAAP, J.(1942), Ann. N.Y. Acad. Sci., 44i

KEMSLEY, W. F. F. (I95I), Ann, Eugen. (Camb.), x6, 316.KIRTLEY, W. R. (I957), Diabetes, 6, 72.KRACHT, J., KRONER, B., VON HOLT, L., and VON HOLT,C. (i957), Ibid., 6, 380.KRALL, L. P., and CAMERINI-DAVALOS, R. (1957), Proc.

Soc. exp. Biol. (N.Y.), 95, 345.LANG, S., and SHERRY, S. (1956), Metabolism, 5, 733.LOUBATIERE, A. (I957), Ann. N.Y. Acad. Sci., 71, 4 (review).MADISON, L. L., and UNGER, R. H. (1958), J. clin. Invest.,

37, 631.MAHLER, R., SHOEMAKER, W. C., and ASHMORE, J. (I958),' Symposium on the Mechanism of Action of Insulin ' (British

Insullin Manufacturers), ,ondon.

MARBLE, A., and CAMERINI-DAVALOS, R. (1957), Ann. N.Y.Acad. Sc., 71, 239.

MARBLE, A. (x958), Med. Clin. N. Amer., 42, 1163.MARTIN, D. B., RENOLD, A. E., and DAGENAIS, Y. M. (1958),Internat. Diab. Fed. Dusseldorf Communic., 35.MARX, R., STICH, W., and EHRHART, H. (1957), Dtsch. med.

Wschr., 82, 1533.McGAVACK, T. H., SEEGERS, W., HAAR, H. O., ENZINGERJ., and ERK, V. O. (r957), Ann. N. Y. Acad. Sci., 71, 268.MILLER, M., CRAIG, J. W., MACKENSIE, M. S DRUCKERW. R., CAMMARN, M., and WOODWAR6, H. (1957),Ibid., 71, 51MIRSKY, A. I., PERISUTTI, G., and GITELSON, W. (1957),Ibid., 71, 103.MOORHOUSE, J. A., KARK, R. M., and GELLMAN, D. D.

(1957), Ibid., 71, 97.PFEIFFER, E. F. (i957), . Endocr., 15, xlviii.PFEIFFER, E. F., STEIGERWALD, H., SANDRITTER W.,BANDER, A., MAGER, A., BECKER, U., and RETIENE, K.

(1957), Med. Klin., 52, 968.PFEIFFER, E. F., STEIGERWALD, H., SANDRITTER, W.

BANDER, A., MAGER, A., BECKER, U., and RETIENE, K.(1957), Dtsch. med. Wschr., 82, x568.

PFEIFFER, E. F., RENOLD, A. E., MARTIN, D. B.DAGENAIS, Y. M., MEAKIN, J. W., NELSON, D. H.SHOEMAKER, G., and THORN, G. W. (1958), Internat.Diab. Fed. Dusseldorf Communic, 75.PHEMISTER, J. C. (1957), Brit. med. J., i, I99.POZZA, G., GALANSINO, G., FOA, P. P. (1956), Proc. Soc. exp.Biol. (N.Y.), 93, 539.PURNELL, R., ARAI, Y., PRATT, E., HLAD, C., and ELRICK,H. (1956), Metabolism, 5, 778.RECANT, L., and FISCHER, G. L. (I957), . clin. Invest., 36, 922.REID, J., MAcDOUGALL, A. I., and ANDREWS, M. M. (95S7),Brit. med. J., ii, 1071.RENOLD, A. E., WINEGRAD, A., FROESCH, E. R., andTHORN, G. W. (1956), Metabolism, 5, 757.RENOLD, A. E., MARTIN, D. B., BOSHELL, B. R., and

THORN, G. W. (I957), Ann. N.Y. Acad. Sci., 71, 71.SHERRY, S., ZEFFREN, J. L., BRAVERMAN, A. E., and

DREY, N. (I957), Ibid., 71, 249.SIREK, A., SIREK, O., MONKHOUSE, F. C., LOGOTHE-

TOPOULOS, J., and BEST, C. H. (1958), Internat. Diab.Fed. Dusseldorf Communic., I89.

STOWERS, J. M., MAHLER, R. F., and HUNTER, R. B(1958), Lancet, i, 278.

SUGAR, S. J. N. (I957), Ann. N. Y. Acad. Sci., 71, 256.TARDING, F., and SCHAMBYE, P. (1958), Endokrinologie, 36,

222.

UNGER, G., FREEDMAN, L., and SHAPIRO, S. L. (1957),Proc. Soc. exp. Biol. (N.Y.), 95, 190.UNGER, R. H., and MADISON, L. L. (I958), Diabetes, 7, 445.VAUGHAN, M. (1956), Science, 123, 885.VAUGHAN, M. (i957), Ann. N.Y. Acad. Sci., 71, II2.VON HOLT, C., VON HOLT, L., and KRONER, B. (1956).

Naturwiss., 43, 162.VON HOLT, C., VON HOLT, L., KRACHT, J., KRONER, B.,

and KUHNAU, J. (1957), Science, 125, 735.WEAVER, J. A., PROUT, T. E., SCOTT, G. W., and ASPER,

S. P. (1958), Brit. med. J., i, 425.WEBER, G., and CANTERO, A. (1958), Metabolism, 7, 333.WALKER, G., SLATER, J. D. H., WESTLAKE, E. K., and

NABARRO, J. D. N. (1957), Brit. med. J., ii, 323.WICK, A. N., BRITTON, B., and GRABOWSKI, R. (1956),

Metabolism, 5, 739.WILLIAMS, R. H., and TUCKER, B. W. (1956), Ibid., 5, 8oi.WILLIAMS, R. H., TYBERGHEIN, J. M., HYDE, P., and

NEILSON, R. L. (1957), Ibid., 6, 311.WRENSHALL, G. A., and BEST, C. H. (1956), Canad. med.

Ass. 7., 74, 968.

References to footnotesCAHILL, G. F., HASTINGS, A. B., and ASHMORE, J. (I957),

Diabetes, 6, 26.FRASER, T. C. R., JOPLIN, G. F., and VALLANCE-OWEN, J.

(1959), Personal communication.LUNDBAEK, K., NIELSEN, K., and RAFAELSEN, O. J.

(1958), Lancet, i, 1036.RECANT, L., and FISCHER, G. L. (1957), Ann. N.Y. Acad.

Sci., 7I, 62.30BEL, G. W., RODRIGUEZ-INIGO, J., MORTON, J. V.,

and LEVINE, R. (x958), Metabolism, 7, 222.

Protected by copyright.

on Decem

ber 29, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.35.403.248 on 1 May 1959. D

ownloaded from

June 1959 MILLER: Peripheral Tuberculous Lymtphadenitis in Childhood 353

sufficient to prevent softening of glands, even whenit is continued for more than a year. Originallystreptomycin and P.A.S. were used, and laterI.N.H. and P.A.S. The latter combination can begiven for long periods without difficulty. Theeffect of giving chemotherapy to children withfirm glands recently infected is interesting, there isan immediate response which is proportional tothe original size of the glands. Thus in a febrilechild with an acute reaction, and a large mass ofglands with much peridenitis, the temperaturefalls rapidly, and the swelling shrinks, often re-solving into two or more discrete glands. In achild with a small gland and little change ingeneral health or bodily reaction the change iscorrespondingly less. But all children improve inhealth and well-being and all recently infectedglands show some reduction in size when firsttreated. This improvement continues for aboutthree to four weeks and then stops, leaving a firm,movable, painless gland. If the chemotherapy iscontinued, the gland may remain in this state formany months but those we have observed havesoftened ultimately. Once softening occurs, skininvolvement is inevitable and then the chance ofachieving a good cosmetic result has largely gone.

In looking after a child with recent tuberculousinfection of glands I therefore use chemotherapyand continue as long as necessary until the firstsign of secondary enlargement of the gland occurs.Then I advise removal of the whole gland and itsimmediate neighbours by careful local dissection.No attempt is made to remove all infected glands,for infection has spread far down the lymphaticchain, but great care is taken with the final suturingof the wound. Chemotherapy is continued for atleast a year after the infection. The prolongedchemotherapy has two purposes: to help to avoidfurther local abscesses and to prevent haemato-genous lesions in other organs. Chemotherapy-isjust as necessary when the lymph glands are super-ficial and visible as it is when the- infection ispulmonary or mesenteric.

Streptomycin (45 mg. per kilo, 20 mg. per i lb.body weight, with a maximum dose of i g. per day)and I.N.H. (20 mg. per kilo, maximum 400 mg.)Yareused to cover the operation and until the wound is

firmly healed. The daily injection of streptomycinis then stopped and P.A.S. (125 mg. per i lb.,300 mg. per kilo) is given instead and continuedfor at least a year.

If the gla.nds are soft when first seen, the situa-tion is not so favourable, but, whether infection isrecent or old, the general problem is the same andchemotherapy should be given as described above.When there is calcium present and infection hasbeen present for a considerable time there is notthe same risk of new haematogenous spread butI.N.H. and P.A.S. should be given for six months.The basic trouble when the gland has softened

is that it cannot be removed as a whole, and ifcaseous material is left at the site of the abscessthen there is always a risk of subsequent abscessformation. Unless the contents of the gland areexceptionally fluid, mere incision is not likely toremove all caseous material; for this reason othermethods, such as curettage or expression, havebeen used, and more recently Lincoln (1957) hasused enzyme irrigation to achieve the same purpose.If the caseous material can be removed healingwill follow, but the final appearance of the scardepends upon the amount of skin loss.The early recognition and effective treatment o

tuberculous lymphadenitis in this way can reducethe length of illness, the chance of complicationsand avoid the pigmented scars seen a generationago.

In conclusion, therefore, it can be seen that thesituation with regard to peripheral tuberculouslymphadenitis is changing rapidly, new cases arebecoming less and less common as the sources ofinfection are brought under control, but old casesinfected some time ago continue to present fortreatment. Peripheral tuberculous lymphadenitisis also still to be seen in adults but, as I indicatedat the beginning of this paper, its natural historyseems different from that of tuberculous adenitisin childhood and it may require different treatment.

BIBLIOGRAPHYANASTASIADES, A. A., TSIKOUDAS, E. C., LINCOLN,E. M., and DALY, J. F. (1957), Amer. Rev. Tuberc., 76, s88.MILLER, F. J. W., and CASHMAN, J. M. (1gs5), Lancet, i, 1286.MILLER, F. J. W., and CASHMAN, J M. (I958), Ibid.,MILLER, F. J, W. (I153), Ibid., i, 3.SMITH, C. M. (I9s58), Med. Offr., 99, 45.

Erratum May Issue-We apologize for the illustration appearing in Dr. J. D. H. Slater's article on 'Oral HypoglycaemicDrugs ' which we regret was included in error.