citizens paid attention to the mechanism that wouldfund this piece of benevolent legislation.

Shell shocked congressmenDuring the spring and summer, however, they did.

They noticed that the dollars would be coming not outof general revenues but out of their own pockets. Andthey revolted. They wrote to their congressmen in suchprofusion that the latter were shell shocked. Theymarched on Washington. They banged on their repre-sentatives' cars. The chairman of the ways and meanscommittee was chased down the street in Chicago byirate elderly citizens. So congress took fright, itsmembers being up for re-election every two years, andvoted in October to repeal the whole bill. At first thesenate demurred. It tried to work out a compromise topreserve at least some of the benefits. The administra-tion might have saved the bill but sat on the sidelines.The health secretary made some feeble gestures ofsupport but was silenced by the budget office. Theissue came to the reconciliation committee, where thecongressmen would not budge. In November theentire law was dumped into oblivion.Among the immediate consequences is some con-

fusion about what taxes need to be paid this year andhow the tax forms are to be filled in. Some of thecollected money was to be used to help the complicatedprocess of balancing next year's federal budget. SomeMedicare beneficiaries, having come to rely on the newlaw, may find themselves without cover or subject tobills for being too long in hospital. Supplementary"Medigap" health policies may rise from $50 to $60-80a month. The states will need to pay more for the healthcare of the uninsured elderly and for certain federallymandated benefits. And observers commented thatthis repeal was the first retreat in a long sequence of

expensive entitlement programmes, predicting thatcongress may think twice before trying again.

Indeed some people thought that more congressmenshould study the bills they are voting for, this beingwhat they are paid for, especially as many of themadmitted that they never quite understood what thebenefits were and who would pay for them. Congress,and even society at large, may have to face up to thereality that somebody has to pay for new programmesand that somehow somebody has to find the money.One approach, a selective tax on a restricted group, isclearly not the way, as shown by the angry protests ofthe bruised Medicare recipients.

But others spoke about justice between generations,painting an unflattering picture of the wealthy elderlynot willing to pay for their less fortunate fellows andleaving the burden to the working young. Theysuggested that the image of the poor widow living oncat food may rapidly give way to that of selfish retiredfolk feasting on crab food and playing volley ball inFlorida. They also predicted that young workerscurrently paying taxes and having large social securitycontributions taken out of their pay cheques will, onaccount of inflation, by no means enjoy the sameaffluence when the enormous baby boom generation inits turn reaches retirement age in the next century.

There also remains the issue of how to providemedical care for the 37 million Americans withouthealth insurance, including the 30% elderly peoplewho are too poor to afford private insurance. Manyhope that this will be achieved through a pluralisticsystem, provided largely at local level. But a strongcontingent still favours a universal health system.Unfazed by the failures of centralised planning ineastern Europe, they dream of past and future gloriousrevolutions and look to a monolithic system run by thebureaucrats in Washington.

Crosshouse Hospital,Kilmarnock KA2 OBEJ Niall MacPherson,FRCP, consultant

Trinity College and StJames' Hospital, Dublin,EireJohn Feely, FRCP, professorand consultant

BrAfedj 1990;300:731-6

New Drugs

Insulin

J Niall MacPherson, John Feely

Since insulins were reviewed in Today's Drugs in 1983,human insulins have become the most commonlyprescribed preparations. Their advent has stimulatedfurther assessment of insulin pharmacokinetics and ofinsulin regimens for achieving good control ofdiabetes.Intense controversy has been occasioned recently bythe suggestion that the counterregulatory hormoneresponse to hypoglycaemia may be less in patientstaking human insulin.

Research is being directed to the feasibility of usingalterations in the structure of the insulin molecule toproduce preparations with therapeutic advantagesover existing insulins. Disposable plastic syringes andneedles and blood glucose concentration testing stripshave become prescribable by general practitioners inthe United Kingdom, and pen devices are being usedmore for subcutaneous insulin administration. Theremaining uncertainties about whether near normo-glycaemia can prevent complications in type I diabeteshave stimulated the long term study phase of thediabetes control and complications trial. The useof insulin for non-insulin dependent diabetics hasattracted increased interest, and the United Kingdomprospective diabetes study will include an assessmentof the effect of this treatment on outcome. Theimportance of educating diabetic patients about their

self care and the complexities of this education processhave also been increasingly recognised.

Insulin preparationsFORMULATIONS

The table summarises currently available insulinpreparations. Soluble insulins consist of insulin insimple solution and are therefore absorbed rapidlywhen injected subcutaneously. They are the onlyinsulins available for intravenous administration. Theinsulin zinc suspensions use the relative insolubility ofinsulin combined with zinc in acetate buffer to retardabsorption from the injection site. As amorphoussuspensions (semilente insulins) have small particlesthey are absorbed rapidly and have a duration of actionsomewhat longer than that of soluble insulin, whereascrystalline suspensions (ultralente insulins) have largerparticles, are more slowly absorbed, and have aduration of action of more than 24 hours. Lenteinsulins comprise a 30% amorphous 70% crystallinemixture. Isophane (NPH) and protamine zinc (PZI)insulins use the relative insolubility of a combination ofprotamine zinc and insulin to retard insulin absorption.

Soluble and isophane insulins may be mixed in thesyringe without the quick action being lost, whereas

BMJ VOLUME 300 17 MARCH 1990 731

on 18 July 2022 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.300.6726.731 on 17 M

arch 1990. Dow

nloaded from

Current/vl available insulins

Type and species Proprietary name

Short actltng neuitral soluible insulinsHuman Actrapid (,pyr)

Human \ closulin (emp)Humulin S (prb)

Porcine VelosulinBovine Hypurin Neutral

Intermediating actingIsophane (NPH) insulins:Human Human Insulatard (emp)

Protaphane (pvr)Humulin I (prb)

Porcine InsulatardBovine Hx purin Isophane

Biphasic insulins:Human Humulins INMl, M12. M3, and M4

10 t) soluble 90% NPH to 40"osoluble 60% NPH) (prb)

Penmix 30/70 (pyr) (30% solubleActraphanc (pyr) 70% NPH)Human Miixtard (emp)Human Initard (emp)(50% soluble 50% NPH)

Porcine MixtardInitard

Mlixed specics Rapitard MC ('25(& porcine soluble75".'s bovine IZS)

Insulin zinc suspensions (IZS):Porcine Semitard MCamorphous

Human lente(30%z,amorphous700s Monotard (psr)crystalline) Humulin Lente (prb

L ong actingInsulin zinc suspensions (IZS):Mixed species L.entard M(3C porcine amorphous,

lente bovine crystalline)Bovine lente Hwpurin LcnteHuman

ultralente Human Ultratard (pyr)(IZS Humulin ZN (prbcrNystalline)

Bovine Hxypurin Protamine Zincprotaminezinc Insulin(PZI)

needles for this purpose have been prescribable.Separate disposable needles are to become prescribable

Peak shortlv. Instructions from the DHSS and Nationalactv tv Duration(hours) (hours) Pharmaceutical Association have been that needle and

syringe should be used once only. Extensive studieshave shown, however, that the reuse of disposableinsulin syringes and needles by the same person is safe.

2-5 5-8 The British Diabetic Association has advised thatplastic syringes may be reused up to five times,although if patients wish to use their syringes only oncethey are free to do so.t A needle clipping device toremove needles from their hubs is prescribable, which

4-12 12-24 ensures their safe disposal.

PEN DEVICES

Increasing numbers of patients are using pen devicesin place of insulin syringes for injection. The NovoPenI is a robust device like a fountain pen and uses

3-8 12 24 cartridges of insulin, Human Actrapid Penfills, that38 12-24 can deliver measured doses of insulin in two unit

increments. Studies have shown a high level of patientacceptance of the device, which is particularly suitablefor facilitating regimens incorporating Actrapidinjections before each of the three main meals. The

5-10 12-16 cartridged insulin, but not the pen or its specialneedles, is available on prescription. Other currentlyavailable pen devices are Penject, which incorporates adisposable insulin syringe, and Autopen and Accupen.

6-14 16-22 Further pen devices becoming available are the Insu-ject pen (for Human Velosulin) and the Insuject X pen(for Human Insulatard, Mixtard, and Initard formula-

6-14 18-30 tions, all provided in cartridges). NovoPen II nowprovides a device for injecting Protaphane and Penmix30/70 formulations.

8-24 24-28

10-20 20-36

emp enzymatic modification of porcine insulini.prb=proinsulin recombinant bacteriol.pyr precursor veast recombinant.

mixing soluble insulin with insulin zinc suspensionsleads to some loss of quick action, probably by solubleinsulin interacting with excess zinc. Mixing solublewith protamine zinc insulin also gives rise to loss ofsome quick action. Biphasic insulins are marketedas mixtures of soluble and isophane (Mixtard, Actra-phane Initard, Penmix 30/70, and Humulins M1-M4)or a mixture of 25% porcine soluble with 75% bovineultralente (Rapitard MC). All insulins are marketed inUIOO strength (100 units/ml). Bovine insulins remainavailable as the Hypurin series of neutral, soluble,isophane, lente, and protamine zinc insulins.

DURATION OF ACTION

For clinical use insulins have been classified tradi-tionally into short, intermediate, and long acting. Therate of absorption of the same insulin, however, showsconsiderable variation and decreases with the size ofthe dose. Thus large doses of "intermediate" actinginsulins can have a long duration of action. Insulinabsorption tends to be fastest from the abdomen andslowest from the resting thigh. Antibodies to insulincan also affect absorption. The classification in thetable should therefore be interpreted in this light.

Administration of insulinINSULIN SYRINGES

Most diabetic patients use disposable plastic insulinsyringes available in 0-5 ml (up to 50 units) and 1 ml(up to 100 units) sizes for subcutaneous injection.Since September 1987 combinations of syringe and

JET INJECTION

Jet injectors are high pressure devices that ejectinsulin through a fine nozzle. The high velocity thusgenerated permits insulin to penetrate the skin withouta needle being used. Jet injectors appear to be able todeliver accurate and reliable amounts of insulin tosubcutaneous tissue. The greater dispersion obtainedgives more rapid absorption of short and intermediateacting insulins and consequently reduces the totalduration of action. Although the initial pain of injec-tion may be reduced, delayed pain and bleeding maybe greater. Denaturation of some insulin by the jet,and consequently increased immunogenicity, is atheoretical possibility. Despite having been availablefor some years jet injectors have not yet achieved wideusage.

Insulin speciesAs recombinant DNA technology has become the

usual mode of insulin manufacture most diabetics inthe UK are now treated with human insulins. 4 Humaninsulin differs from bovine in that the alanine ofA8 andvaline of A10 of the bovine A chain are threonine andisoleucine respectively in the human A chain, and the Cterminal (B30) amino acid of the B chain (which isalanine in the cow) is threonine in the human molecule.Human and porcine insulins differ only in that the B30amino acid is alanine in the pig.To date, human insulin has been produced on an

industrial scale by four different processes. The Novo/Nordisk company has previously produced humaninsulin by enzymatic modification of porcine insulin(emp) but the Novo division now uses recombinantDNA technology in yeasts to produce human insulin(pyr) derived from proinsulin. Lilly originally madeinsulin from the combination of separate A and Bchains produced by recombinant DNA technologyusing fermentation in Escherichia coli with subsequentchemical combination of the chains (crb). This process

BMJ VOLUME 300 17 MARCH 1990)732

on 18 July 2022 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.300.6726.731 on 17 M

arch 1990. Dow

nloaded from

has now been supplanted by production from pro-insulin in E coli (prb).Human insulin is slightly more soluble than porcine

and bovine insulins, and a more rapid subcutaneousabsorption, particularly of long acting formulations,might therefore be predicted. Pharmacokinetic studiesof human ultralente insulin confirm that it is morerapidly absorbed than the bovine formulation, hasearlier and more definite peak insulin levels, but has asufficiently long duration of action to make humanultralente a suitable preparation to provide basalinsulin requirements by injection once a day.' Moststudies of intermediate acting isophane and lentehuman insulin formulations and of human solubleinsulins show that they have a slightly shorteraction than porcine preparations when subcutaneouslyinjected.

The controversy about human insulin andhypoglycaemiaFrom 1985 to 1989 the proportion of British diabetic

patients treated with human insulin increased fromaround 6% to more than 75%.4 As with previouschanges in the 1970s to highly purified (mono-component) insulins and in the early 1980s to U100insulin, some patients reported changes in theirwarning symptoms of hypoglycaemia on transfer tohuman insulin.6 The possibility that human and animalinsulins might cause different counterregulatoryhormone responses to hypoglycaemia, however, andmedia coverage of the further suggestion that change tohuman insulin had been associated with an increase insudden unexplained deaths, led to the current intenseinterest in possible relations between human insulinand hypoglycaemic unawareness."

Several cases of altered experience of hypoglycaemiaafter transfer to human insulin can be explained fairlyreadily. Bovine insulin is more immunogenic thanhuman or porcine insulin, and a lower antibody titre ontransfer to human insulin may alter dose requirementsand duration of action. The change to human insulinmay be made when the regimen is altered to improvecontrol. Such improvement in glycaemic controlhas been associated with a lowered glucose thresholdfor the release of counterregulatory hormones,6 witha consequent delay in experiencing adrenergicsymptoms. The more rapid absorption of humaninsulin may also on occasion give rise to quickerchanges in blood glucose concentrations, thus alteringsymptoms of hypoglycaemia.These factors operating when there has been change

from bovine insulin or an alteration in insulin regimenmay account for much of the altered symptoms ofhypoglycaemia on transfer to human insulin. Moredifficult to explain are suggestions of altered hypo-glycaemic awareness after transfer from highly purifiedporcine insulins to the same regimens of humaninsulin, when the altered pharmacokinetics might beexpected to cause only minor differences. Severalcrossover trials have not shown any difference in hypo-glycaemic experiences between human and porcineinsulin,9 but other studies suggest a high incidence ofhypoglycaemic unawareness on transfer to humaninsulin."'" These, however, have been criticised forinappropriate questionnaire design and implausiblyhigh rates of severe hypoglycaemia.7 Surveys of experi-ence of hypoglycaemia after change to human insulinare similarly conflicting. A British Diabetic Associa-tion questionnaire showed that 24 out of 158 patientschanging to human insulin experienced less hypo-glycaemic warning,'2 whereas in The NetherlandsHeine and van de Veen recorded a steady reduction inthe number of patients with severe hypoglycaemiaadmitted to hospitals during 1983-8, when the use of

human insulin was increasing." About a third ofdiabetic patients usually lose adrenergic symptoms ofhypoglycaemia during the first 20 years. Gale haspointed out how this could lead to attribution bias,- so1500 of 100 000 patients a year can be expected toreport altered symptoms even if treatment is notchanged.

Studies of the counterregulatory hormone responseto experimental hypoglycaemia in normal subjects aresimilarly conflicting: some show no difference betweeninsulins9 and others show reduced responses of nor-adrenaline or altered symptoms when human insulin istaken. '4 Although the trigger to counterregulatoryhormone response is hypoglycaemia, insulin affects therelease of noradrenaline independently of its effectson blood glucose concentrations. This leads to sugges-tions that porcine insulin (which is slightly morelipophilic than human insulin) could have a quickeraction because it penetrates the blood-brain barriermore quickly. The relevance of these mechanisms tothe hypoglycaemia experienced by diabetic patients isnot known.While controversy over human insulin and hypo-

glycaemia continues, certain policies seem to be appro-priate. Newly diagnosed patients should be givenhuman insulin, and those who have been taking humaninsulin since their diagnosis are very unlikely to gainbenefit from a transfer to animal insulin. Tight controlof diabetes is associated with increased hypoglycaemiaand reduction in warning symptoms whatever insulinis used, so treatment objectives should not be un-realistically high. During the past six years manypatients have been successfully transferred frombovine to human insulins, so those in the UnitedKingdom still taking bovine insulins are increasinglylikely to be a small group of patients with long standingdiabetes, the course of which has been particularlyfavourable. The only indications for transfer frombovine insulins- lipoatrophy or allergy to insulin -areunlikely to apply to these patients, and as long assuitable bovine preparations remain available doctorsmay deem it wise to avoid unnecessary change. Whenchange is required an initial dose reduction of 10%has been suggested for patients taking less than 0 9units/kg and a 25% reduction for those taking higherdoses. Return to original animal insulins shouldbe considered for patients who have experiencedproblems on transfer from animal to human insulinthat cannot be explained by changes in regimen orcontrol and for patients who have lost confidence inhuman insulin.

Indications for and objectives of treatment withinsulinBy definition, insulin is indicated for type I (insulin

dependent) diabetes. Regimens sufficient to avoidprimary diabetic manifestations of thirst, polyuria,weight loss, and ketosis are easily achieved. The aimfor most patients, however, is to normalise theirmetabolism sufficiently to prevent the long termcomplications of diabetes developing. A wealth ofdata and results of recent trials to maintain nearnormoglycaemia for relatively short periods provideencouragement that this may prove possible. Currentevidence is not conclusive, however, which justifies theinvestment in the currently running diabetes controland complications trial of intensified versus conven-tional glycaemic control.' '"6This follows up a primaryprevention group as well as patients with early retino-pathy and assigns adolescents as well as adults forrandomisation. The numbers enrolled give hopefor clear conclusions. Meanwhile, for young type Ipatients it is reasonable to aim for the best glycaemiccontrol possible within the following constraints.

BMJ VOLUME 300 17 MARCH 1990 733

on 18 July 2022 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.300.6726.731 on 17 M

arch 1990. Dow

nloaded from

Sudden near normoglycaemia after mediocre control ofdiabetes can cause acute deterioration in retinopathy,so monitoring optic fundi is particularly importantduring acute improvement in control. Near normo-glycaemia is associated with a significant incidence ofhypoglycaemic episodes,'6 and for most patients canonly be achieved at the expense of considerable timeand effort. Near normal glycated haemoglobin valuesachieved in clinical trials are not representative of thegeneral diabetic population, in whom values severalstandard deviations above normal are the norm. Formany patients, achievements in diabetic control areperhaps most realistically measured against improve-ment over this norm rather than an unrealistic goal ofnormality.

Insulin has a place in the management of some typeII (non-insulin dependent) diabetic patients. Thedominance of hyperglycaemia as the major risk factordetermining outcome is perhaps more controversial fortype II than type I diabetes. In populations with earlyonset of type II diabetes (such as Pima Indians andAsians in Britain) the natural history and the roleof hyperglycaemia in determining outcome may besimilar to that in type I patients. In European popula-tions presenting with type II diabetes in middle or laterlife, however, the outcome is heavily influenced byan excess of cardiovascular disease, which may asplausibly be related to the interaction of hyper-glycaemia with hypertension, hyperlipidaemia, hyper-insulinaemia, and a postulated genetic associationbetween type II diabetes and cardiovascular disease asto hyperglycaemia alone. The only major prospectivetrial of insulin treatment in type II diabetes, theuniversity group diabetes programme showed nobetter outcome with insulin than with diet alone,despite better glycaemic control with insulin.'7 TheUnited Kingdom prospective diabetes study hasenrolled a substantial cohort of type II subjects treatedwith insulin'8 but has not as yet produced any reportson outcome. Pending these results it is reasonable toaim for as good glycaemic control as possible in type IIdiabetic patients who are not old and whose lifeexpectancy is not limited by other medical conditions.Most type II patients are overweight, and strenuousefforts should be made to achieve good control throughdietary measures before considering the use of insulin.Underweight or normal weight patients receivingmaximum oral treatment but who have poor glycaemiccontrol are severely insulin deficient and should betreated with insulin.

For insulin dependent and non-insulin dependentdiabetic patients the control of hypertension, smoking,and hyperlipidaemia are increasingly recognised asbeing important to prevent macrovascular disease.Control of hypertension also substantially slows therate of progression of diabetic nephropathy.

Insulin regimensIn normal subjects insulin is secreted at two rates: a

basal rate during fasting to exert an inhibitory controlon the catabolic processes of glycogenolysis, gluco-neogenesis, lipolysis, and the breakdown of proteins,and a rapid rate in response to meals to promote thestorage of absorbed fuels. Insulin regimens shouldmimic this pattern to achieve good control.

INTENSIFIED CONTROL REGIMENS

In continuous subcutaneous insulin infusion abattery powered pump infuses soluble insulin througha subcutaneous cannula to provide the basal rate andboluses are given before meals through the pump. Inone form of intensified conventional treatment solubleinsulin is given three times a day before meals, oftenusing a pen device for convenience of injection; the

basal insulin concentration is provided by an injectionof ultralente insulin, usually given at night beforebed, as this preparation's prolonged absorption timeprovides fairly consistent background insulin concen-trations once a steady state has been reached. Alterna-tively, an intermediate acting insulin given at bedtimecan be used to provide overnight basal insulin concen-trations. If patients are carefully selected, motivation ismaintained, and intense supervision is applied thenboth these techniques can produce near normal glucoseand glycated haemoglobin concentrations, though con-tinuous subcutaneous insulin infusion tends to giveslightly better results.'9 2 It is equally clear, however,that if these conditions do not apply continuoussubcutaneous insulin infusion or frequent injections donot give better results than less intense regimens.Patients most suitable for these approaches to selfmanagement are likely to be those who have alreadyshown high motivation and the ability to obtaingood glycaemic control with conventional regimens.Patients with brittle diabetes or who on psychologicaltesting perceive the pump or pen technology as ex-ternalising the responsibility for their diabetes areunlikely to do well. Continuous subcutaneous insulininfusion carries the possible disadvantage that if thepump fails the onset of ketoacidosis may be more rapidand more likely to be associated with dangeroushyperkalaemia because there is no continuing depot ofsubcutaneous insulin. For these reasons continuoussubcutaneous insulin infusion has not achieved wide-spread use in the United Kingdom. By contrast, theuse of frequent injection regimens using pen devices isincreasing as, even when good diabetic control is notachieved, they may help patients achieve flexibility ofmealtimes and activities and the psychological benefitof the feeling that they have more control over theirdiabetes.

TWICE DAILY INSULIN REGIMENS

The insulin regimen used by most type I diabeticpatients remains the combination of short and inter-mediate acting insulin injected twice daily, beforebreakfast and before the evening meal. The morningshort acting insulin is intended to produce a rise ininsulin concentration after breakfast, and the inter-mediate acting insulin injected in the morning reachesits peak action in the afternoon. Insulin action after theevening meal and during the night is similarly providedby the evening injection of short and intermediateacting insulins respectively. Soluble and isophane orlente formulations are used. A common difficulty withthese regimens is morning fasting hyperglycaemiabecause the duration of action of the evening inter-mediate acting insulin is insufficient.7' Increasing thedose of evening intermediate acting insulin tends toproduce hypoglycaemia (often asymptomatic) around3 am. Delaying the intermediate insulin injection tobefore bedtime may help. Substituting ultralenteformulations for the intermediate acting insulin canalso be tried, but may produce more hypoglycaemia inthe hours before waking.

In the adjustment of twice daily regimens patientshave been asked traditionally to adjust the dose ofmedium acting insulin given before breakfast accord-ing to the blood glucose concentration before theevening meal and the evening dose according to theblood glucose concentration before breakfast. Thevariability of absorption of medium acting insulinsis such, however, that these adjustments can onlylogically be made on the basis of a trend in bloodglucose concentrations for several days. The lack ofneed for fine adjustment of dose in these regimensjustifies the use of premixed biphasic soluble isophanecombinations and adjusting the mix and dose to longterm trends in blood glucose concentrations. The

BMJ VOLUME 300 17 MARCH 1990734

on 18 July 2022 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.300.6726.731 on 17 M

arch 1990. Dow

nloaded from

pharmacokinetics of subcutaneous soluble insulininjections are more favourable for rapid adjustments indose and provide further justification for the use ofsoluble insulin three times a dav and intermediateacting insulin at bedtime by patients seeking flexibilityand good glycaemic control.

INSULIN REGIMENS FOR NON-INSULIN DEPENDENTDIABETICS

When insulin is used to treat non-insulin dependentdiabetic patients" regimens for patients with dominantinsulin deficiency need not differ from these used fortype I subjects, and twice daily biphasic insulins area reasonable choice. Insulin requirements will beincreased, often substantially, in obese patients. Insubjects with less pronounced insulin deficiency therelief of fasting hyperglycaemia caused by overnightinsulin deficiency will probably improve endogenousinsulin response to meals, thereby also reducingpostprandial glycaemic excursions. Regimens using asingle daily injection of bedtime ultralente or isophaneinsulins at a dose adjusted to the fasting morning bloodglucose concentration have therefore been proposedfor those with moderate insulin deficiency.

Using the insulinsUnlike most other drug treatments, achieving good

metabolic control of diabetes is not simply a matter ofadjusting insulin dose to response. This is because ofthe very imprecise relation between effect and the doseof subcutaneous insulin and because many variablesother than insulin affect the control of diabetes. Innormal physiology insulin is secreted in a pulsatilemanner into the portal circulation, and its secretionis finely regulated from minute to minute by stimula-tion by gut derived factors and the prevailing bloodglucose concentration, which is in turn influenced bythe processes of feeding, fasting, exercise, prevailinglevel of insulin sensitivity, and counterregulatoryhormone levels.

Current replacement treatment is an attempt toimitate this physiological activity with subcutaneousinjections of insulin that have coefficients of variationfor half time of absorption of about 25% in individualpatients. Insulin is then absorbed into the systemic, asopposed to portal, circulation to act on tissues with asensitivity to insulin that varies considerably from dayto day. The whole system is administered by thepatient, for whom treating and monitoring diabetes isonly one of life's many daily tasks. This dominance ofvariability in determining the effect of subcutaneousinsulin injections implies that insulins cannot be usedas "precision tools," that attention to factors other thaninsulin is important in determining diabetic control,and that any regimen producing near normoglycaemiamust carry a certain incidence of hypoglycaemia.Newly diagnosed insulin dependent diabetic patients

presenting with hyperglycaemia but not ketosis maybe introduced to insulin treatment as outpatientsprovided that regular adequate contact with a diabeticspecialist nurse is available. Urgent restoration ofnormoglycaemia is not required for such patients andthe initial insulin requirement may be found empiric-ally over a period of days. Those seeking a guide todetermining initial insulin requirements may find thetables published by Holman and Turner helpful.4 Allpatients who are capable of benefiting should beinstructed in monitoring blood sugar concentrations athome.

For patients already taking insulin who encounterproblems with poor or unstable diabetic control, thebasic practicalities of diabetic self care should beassessed before considering complexities in the insulinregimen. Some recently diagnosed patients may inter-

mittently stop taking insulin either because in theirremission phase they believe they no longer havediabetes, or as a form of denial of the condition, orsimply because they forget an injection, particularlythose using multiple injection regimens. Problems ofthis type will be detected only if a suitably relaxedlistening environment, in which the patients' viewsand feelings may be expressed, is provided. Lipo-hypertrophy caused by repeated injection into thesame site may cause pronounced variability in insulinabsorption and must be looked for. Chaotic lifestyles of teenagers, particularly regarding the times ofsleeping, waking, and eating, may severely disruptcontrol. Conversely, obsessional patients may generatediabetic instability by the overfrequent and excessiveadjustment of insulin dose in response to homemonitoring of blood sugar concentrations.Even when adherence to treatment is good, fasting

hyperglycaemia is a common problem.)' Three concep-tually separate but interacting phenomena probablycontribute to this. The waning of free insulin con-centrations in the early hours of the morning aftertaking intermediate acting insulins in the evening hasalready been referred to. The dawn phenomenon is anincrease in insulin requirement in the few hours beforewakening, which may be caused by a rise in growthhormone concentration among other factors. In theorycontinuous subcutaneous insulin infusion regimensthat increase the basal rate at about 4 am to 8 am mightovercome this problem. When a decrease in sensitivityto insulin is caused by the counterregulatory hormoneresponse to preceding hypoglycaemia the resultanthyperglycaemia is commonly referred to as the Somogyiphenomenon. Hypoglycaemia at about 3 am iscommon with twice daily insulin regimens and can giverise to fasting hyperglycaemia through the Somogyieffect. This phenomenon and other effects of excessinsulin may possibly operate at other times of the day.When insulin requirements exceed one unit/kg/day thepossibility that the insulin dose is too high should beconsidered if recognised causes of high insulin require-ment (such as adolescence, obesity, or pregnancy) areabsent. An increase in the insulin requirement of apreviously well controlled patient may be caused by theonset of hyperthyroidism, which occurs more often intype I diabetics. Decreased requirements may also becaused by hypothyroidism, Addison's disease, or,advancing diabetic nephropathy. Concurrent illness ofany type may increase the insulin requirement.The need to educate insulin dependent patients

about their self care has long been recognised. Thelimitations of simply imparting information, however,have increasingly been apparent,2' and attention isbeing directed to ways of changing the attitudes todiabetes that impede self care. Such techniques startwith patients' existing concepts of their diabetesand encourage their exploration of these concepts.Although opinions about educational methods mayvary, patients should reasonably expect to attend acentre where an active interest in education is main-tained.

Future developmentsThe advent of recombinant DNA technology in

the manufacture of human insulin has resulted inthe availability of effective insulins of lower im-munogenicity but with relatively few obvious thera-peutic advantages over the highly purified porcineinsulins that they replace. The technology createdhas, however, permitted the creation of new aminoacid substitutions that favourably alter the pharmaco-kinetics or pharmacodynamics of the molecules ofwhat have been called "designer insulins."27Commercially available soluble insulin preparations

BMJ VOLUME 300 17 MARCH 1990 735

on 18 July 2022 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.300.6726.731 on 17 M

arch 1990. Dow

nloaded from

exist in the form of hexamers, and the delay inabsorption from subcutaneous tissue is contributed toby the time required for dissociation into dimericor monomeric forms. Substituting aspartate for theB9 and glutamate for the B27 threonine residues ofhuman insulin alters the interaction between insulinmonomers and results in monomeric insulins withnormal receptor binding. Recent studies show quickersubcutaneous absorption of this insulin analogue andmore rapid hypoglycaemic effect, which promisesbetter matching between rises in glycaemia after mealsand insulin concentrations.28The insulin molecule has also been modified to

change the isoelectric point towards neutral valueswhile preserving stability, thus producing insulin that,when injected as a slightly acid solution, crystallisesat neutral pH in the tissues and delays absorption.Such insulins may be able to combine less variableabsorption than current suspensions with long absorp-tion times.'The possibility of insulins with greater activity on

hepatic glucose output than on peripheral glucosedisposal has also been raised. Proinsulin was at onetime a candidate for such a substance but has now beenwithdrawn from clinical trials.

1 Alexander WD, Tattersall R. Plastic insulin syringes: reuse or waste £8m ayear. BrMedj 1988;296:877-8.

Summarises evidence for the safe reuse of syringes, current regulations,and British Diabetic Association advice.

2 American Diabetic Association task force on jet injectors. Position statementon jet injectors. Diabetes Care 1988;11:600-1.

Current status of jet injectors.3 Pickup J. Human insulin. BrMedJ 1986;292:155-7.

Review of human insulin and its clinical use.4 Pickup J. Human insulin-problems with hypoglvcaemia in a few patients.

BrMedJ7 1989;299:991-3.5 Hildebrandt P, Berger A, Voland Aa, Kuhl C. 'I'he subcutaneous absorption of

human and bovine ultralente f'ormulations. DiabeticMedicine 1985;2:355-9.Absorption rate of human ultralente, though faster than that of bovine

ultralente, makes it suitable as the basal insulin preparation for multipleinjection regimens.

6 Anonymous. Transferring diabetic patients to htiman insulin. Lancet 1989;i:762-3.

7 Gale EAM. Hypoglycaermia anid human insulin. Lancet 1989;ii: 1264-6.8 Amicl S, Sherwin RS, Simonson DC, Tamborlane WV. Effect of intensive

insulin therapy or glycaemic thresholds for counter-regulatory hormonerelease. Diabetes 1988;37:901-7.

9 Berger M. Htuman insulin, much ado about hypoglycaemic (un)awareness.Diabetologia 1987;30:829-33.

10 'lTeuscher A, Berger WG. Hypoglycaemia uinawareness in diabetics transferredfrom beef/porcine inisulin to human insulin. Lancet 1989;ii:382-5.

11 Berger W, Kcller U, Honiegger B, Jaeggi E. Warning symptoms of hypo-glycacmia during treatment with human and porcine insulin in diabetesmellitus. Lanzcet 1989;i: 1041-4.

12 Redmond S. Changing to human insulin. Balance 1988;Aug/Sept:66-7.13 Heine RJ, van de Veen FA. Human insulin and hypoglycaemia. Lancet

1990;335:62.14 Heine Rj, van der Heyden EAP, san de Veen FA. Responses to human and

porcine instilin in healthy subjects. Lancet 1989;ii:946-8.15 Diabetes Control and Complications 'I'rial DCCrr) Research Group. Are

continuing studies of mctabolic control and microvascular complications in

inisulin dependent diabetes mellituis jslstified? N Engl J Med 1988;218:246-50.

Review of evidence that good glvcaemic control mav prevent diabeticcomplications, with conclusion that case is not vet proved. References forcontrary view cited.

16 DCCT Research Group. Diabetes control anid complications trial (DCCT).Resuilts of feasibilitv study. Diabetes Care 1987;10:1-19.

Structure of the important DCCT trial. Evidence of increased hypo-gl!'caemia with intensified control.

17 University Group Diabetes Programme. Effects of hypoglycaemic agentson vascular complications in patients with adult onset diabetes. VIII.Evaluation of insulin therapy: final report. Diabetes 1982;31:suppl 5:1-8 1.

Insulin treatment did not improve death rate or outcome of microvasculardisease in type II diabetics.

18 AMulticentre Study. UK prospective study of therapies of maturity onsetdiabetes. I. Effects of diet, sulphonylurea, insulin or biguanidc therapy orfasting plasma glucose and body weight over one year. Diabetologza1983;24:404-1 1.

Structure of UK prospective trial that will include assessment of role ofinsulin treatment on outcome of type II diabetes.

19 Marshall SM, Home PD, Taylor R, Alberti KGMAM. Continuous sub-cutaneous insulin infusion versus injection therapy: a randomized cross-overtrial undcr usual diabetic conditions. Diabetic Medicine 1987;4:521-5.One comparison of continuous subcutaneous insulin infusion and opti-

mised injection treatment showing comparable control. Referenccs to othercontinuous subcutaneous insulin inftusion trials cited.

20 Saurbrey N, Arnold Larsen S, Moller Jensen B, Kuhl C. Comparison ofcontinuous subcutaneous insulin infusion with multiple insulin injectionsusing the NosoPen. Diabetic Medicine 1988;5:150-3.

Trial showing slightly better control with continuous subcutaneousinsulin infusion but patient preference for multiple injections with NosvoPen.Other studies cited.

21 Francis AJ, Home PD, Watford S, Alberti KGMM, Mann N, Reeves WG.Prevalence of morning hyperglycaemia, determinants of fasting bloodglucose concentrations in insulin treated diabetics. Diabetic Medicine1985;2:89-94.

Incidence of fasting hyperglycaemia in type I diabetics receiving currentinsulin regimens.

22 Lauritzen r, Pramming S, Gale EAM, Deckert T, Binder C. Absorption ofisophane (NPH) insulin and its clinical implications. Br Med J 1982;285:159-62.Pronounced variability in absorption of isophane insulin showil and

consequences discussed.23 Tattersall B, Scott AR. When to use insulin in the maturity onset diabetic.

Postgrad MedJ7 1987;63:859-64.Balanced view of the problem of when to use insulin in type II diabetes.

24 Holman RR, Turner RC. A practical guide to basal and prandial insulintherapy. Diabetic Medicine 1985;2:45-53.

Guidelines for insulin doses in patients starting insulin treatment.Description 'f rationale for basal and prandial treatment.

25 Gerich JE. Glucose counter-regulation and its impact on diabetes mellitus.Diabetes 1988;37:1608-17.

Glucose counterregulatory mechanisms and their effect on diabeticcontrol.

26 Anderson RM. The personal meaning of having diabetes. Implications forpatient behaviour and education or kicking the bucket theory. DiabeticMedicine 1986;3:85-9.

Introduction to the concept that self care of diabetics depends on attitudesas well as knowledge and requires appropriate educational approaches.

27 Pickup J. The pursuit of perfect control in diabetes. Better insulin betterdelivered. Br Medj 1988;297:929-3 1.

Description of possibilities in modifying insulin molecule as well as novelways of delivering insulin.

28 Vora JP, Owens DR, Dolben J, et al. Recombinant DNA derived monomericinsulin analogue: comparison with soluble human insulin in normalsubjects. BrMedJ 1988;297:1236-9.Shows more rapid subcutaneous absorption of monomeric insulin

analogue.29 Jorgensen S, Vaag A, Langkiaer L, Hougaard P, Miarkussen J. NovoSol Basal:

pharmacokinetics of a novel soluble long acting insulin analogue. Br Medj1989;299:415-9.Long acting soluble insulin analogue shows less sariability in absorption

than Ultratard.

ANY QUESTIONS

Is there any basis for the belief that stout enhances the quality of breast milk?

A prevalent belief is that beer drinking increases production of breastmilk. Most drugs that increase milk production-for example, dopamineblocking agents like sulpiride and metoclopramide-do so by increasingthe plasma prolactin concentration. Yet the effects of alcohol on prolactinrelease are arguable.' De Rosa and coworkers, however, studied young,non-lactating women to distinguish between the effects of beer from thoseof alcohol on plasma prolactin concentration.2 In the fasting state one litreof beer resulted in a more than twofold increase in the plasma concentra-tion of prolactin within 30 minutes. In contrast, neither a litre of sparklingwater nor a litre of a mixture of alcohol and sparkling water (with the samealcohol content and effervescence as the beer) had any effect on plasmaprolactin concentration. Apparently, beer specifically affected prolactinrelease, perhaps independently of its alcohol content.Whether beer should be unreservedly recommended for lactating

mothers is another matter. A recent American study of 400 mother and

infant pairs suggested that regular alcohol consumption by breast feedingmothers (one standard drink or more a day over three months) wasassociated with a significant fall in the infant's psychomotor developmentscores at 1 year.3 The authors speculatively attributed this to alcoholinduced neuroanatomical damage at a vulnerable period of brain develop-ment. It would seem prudent for doctors to advise caution over the regularuse of alcohol during breast feeding. Nevertheless, if beer does stimulatethe production of prolactin, which is independent of alcohol intake, then itmight be worth while exploring whether alcohol free beer was effective. -A LUCAS, head of infant and child nutrition, MRC Dunn Nutrition Unit,Cambridge1 Ylikahri RH, Huttunen MO, Harkonen M. Effect of alcohol on anterior pituitary secretion of

trophic hormones. Lancet 1976;i: 1353.2 De Rosa G, Corsello SM, Ruffilli MP, Della Casa S, Pasargiklian E. Prolactin secretion after

beer. Lancet 1981;ii:934.3 Little RE, Anderson KWY, Ervin CH, Worthington-Roberts B, Clarren SK. Maternal alcohol use

during breast-feeding and infant mental and motor development at one year. N EnglJ Med1989;321:425-30.

736 BMJ VOLUME 300 17 MARCH 1990

on 18 July 2022 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.300.6726.731 on 17 M

arch 1990. Dow

nloaded from