a systematic review of evidence for the appropriateness of neonatal screening ... › d164 ›...

Journal of Public Health Medicine Vol. 20, No. 3, pp. 331-343Printed in Great Britain

A systematic review of evidence for theappropriateness of neonatal screeningprogrammes for inborn errors of metabolismMargaret J. Thomason, Joanne Lord, Murray D. Bain, Ronald A. Chalmers,Peter Littlejohns, G. Michael Addison, A. Hervey Wilcox andCarol A. Seymour

AbstractBackground Developments in screening technology andincreased understanding of the natural history and treatmentof inborn errors of metabolism (lEMs) have producedpressure to extend neonatal screening programmes. Thisreview aims to assess the evidence for the appropriateness ofsuch programmes.

Methods A formal systematic literature review was con-ducted. Exclusion and inclusion criteria were used to selectpapers for critical appraisal by pairs of reviewers. Standardcriteria were used to assess the appropriateness of neonatalscreening for various lEMs. Site visits were conducted toassess new technologies for newborn screening.

Results A total of 1866 papers were identified and 407systematically selected for full critical appraisal. Publishedevidence confirmed that universal newborn screening forphenylketonuria (PKU) meets all of the screening criteria andjustifies the expense and infrastructure necessary for thecollection and testing of neonatal blood spots. There wasinsufficient evidence in the literature to assess the cost-effectiveness of screening for any other lEMs. There wasreasonable evidence to support inclusion in extendedneonatal screening of four other lEMs: biotinidase deficiency,congenital adrenal hyperplasia (CAH), medium-chain acylCoA dehydrogenase (MCAD) deficiency and glutaric aciduriatype 1 (GA1).

Conclusions Large-scale trials of screening for biotinidase,CAH, MCAD and GA1 should be conducted, with carefulevaluation to establish their clinical effectiveness and cost-effectiveness in practice. Screening for the latter twodisorders would be dependent upon the use of tandemmass spectrometry (tandem MS). The application of tandemMS to newborn screening requires further evaluation. Theextension of neonatal screening programmes to other lEMsis not currently justified.

Keywords: neonatal screening, inborn errors of metabolism,mass spectrometry, literature review

Introduction

The first population-based neonatal screening programmes fordetecting inborn errors of metabolism (lEMs) were introduced

around 30 years ago for phenylketonuria (PKU). Since that timethere has been a substantial increase in the number of identifiedIEMs and although each individual disease is rare the totalincidence is significant. For example, in the United Kingdom,with around 793 000 infants born each year, there are 60-70affected by PKU,1 60 infants with congenital adrenal hyper-plasia (CAH), 50-60 with medium-chain acyl CoA dehydro-genase (MCAD) deficiency and 40-50 with disorders of organicacid metabolism. It has been estimated that these disorderscollectively have an impact on health care comparable with thatof juvenile-onset diabetes.2

More than 99 per cent of newborns in the United Kingdom,and infants in most other developed countries, are currentlytested for PKU and also for congenital hypothyroidism.3'4 Thereis considerable pressure to expand universal newborn screeningprogrammes to cover a broader range of IEMs. This has beenfurther supported by the development of new technologieswhich are applicable to multi-disease screening programmes,for example tandem mass spectrometry (tandem MS) and

'Department of Child Health, St George's Hospital Medical School, LondonSW17 ORE.2Health Care Evaluation Unit, Department of Public Health Sciences, St George'sHospital Medical School, London SW17 ORE.3Royal Manchester Children's Hospital, Manchester M27 1HA.4St Helier Hospital, Carshalton, Surrey SM5 1AA.

'Department of Cardiological Sciences, St George's Hospital Medical School,London SW17 ORE.

Margaret Thomason, Health Services Research Fellow1

Joanne Lord, Lecturer in Health Economics2

Murray Bain, Consultant Paediatrician'

Ronald Chalmers, Professor of Paediatric Metabolism'

Peter Littlejohns, Professor of Public Health2

Michael Addison, Consultant Chemical Pathologist3

Hervey Wilcox, Consultant Chemical Pathologist4

Carol Seymour, Professor of Clinical Biochemistry and Metabolism5

Address correspondence to Dr Margaret Thomason, Health Care EvaluationUnit, Department of Public Health Sciences, St George's Hospital MedicalSchool, Cranmer Terrace, London SW17 ORE.

© Oxford University Press 1998

332 JOURNAL OF PUBLIC HEALTH MEDICINE

automated immunoassay. Before the introduction of any suchscreening programme there should be good evidence that theadditional screening and consequent treatment costs, and thepotential anxiety to parents, are outweighed by the benefits ofhealth improvements and savings in health care. The fact that awide range of metabolic disorders can be detected by using newtechnologies does not necessarily mean that screening for thesedisorders should be undertaken. The longer-term implicationsfor individuals, the health service and society should also beconsidered. For some IEM disorders it is not even clear thatearly detection offers any benefit to the child.5

A health technology assessment project was undertaken inthe United Kingdom to assess current newborn screeningprogrammes, to evaluate the available evidence for theirexpansion and to make recommendations for future develop-ments. Two fundamental questions were addressed: (1) WhichIEMs should be screened for? (2) What technologies should beused for screening?

A consortium was established to review the technical,clinical and economic implications of screening. The projectteam consisted of a combination of the disciplines of clinicalbiochemistry, metabolic medicine (both adult and paediatric),public health and health economics. Two members of this teamwere also experts on the automation of technologies and onmass spectrometry. Although the aim was to provide advice tothe National Health Service the literature review was interna-tional and the results are pertinent to those countries withcomparable incidence rates of IEMs to those in the UnitedKingdom. To avoid compromising the high rate of coverage ofcurrent newborn screening programmes the assumption wasmade that the infrastructures for sample collection would not besubstantially altered.

Methods

Three methods of data collection were used: a formalsystematic review of the literature was conducted, a ques-tionnaire was sent to all newborn screening laboratories in theUnited Kingdom to enquire about current practice and greyliterature, and visits were made in the United Kingdom, UnitedStates and Finland to assess potential new technologies fornewborn screening.

For the purpose of the systematic review a classicaldefinition of an IEM was taken, i.e. a monogenic diseaseresulting in deficient activity in a single enzyme in a pathway ofintermediary metabolism. Disorders thus considered were thecatabolic and synthetic pathways of carbohydrates, amino acids,organic acids, fatty acids, purines, pyrimidines, porphyrins,steroids, lipids and bile acids, and secondly of the processesinvolved in the uptake, synthesis and utilization of the essentialcofactors for enzymes in these pathways, including trace metalsand vitamins; also included were disorders of lyosomal enzymes.The categories of IEM included in the review were:

(1) phenylketonuria;

(2) amino acidopathies;(3) disorders of carbohydrate metabolism;(4) disorders of organic acid metabolism;(5) fatty acid oxidation defects;(6) disorders of adrenal steroidogenesis;(7) lipoprotein disorders;(8) peroxisomal disorders;(9) disorders of the urea cycle;

(10) respiratory chain/tricarboxylic acid cycle disorders;(11) trace metal disorders;(12) purine-pyrimidine disorders;(13) lysosomal disorders.

Other inherited disorders, for example of membrane transport,of connective tissue, of blood and of blood-forming tissues, ofthe defence and immune systems, and of muscle and skin wereexcluded. Congenital hypothyroidism, which can be caused bymetabolic defects of thyroxine synthesis, was not considered.

The search methods and selection criteria are shown in Fig.1. A total of 407 papers were selected for critical appraisal. Foreach of the 13 categories of IEM at least two subject expertsread and critically appraised all of the selected papers using anagreed checklist, which included study population, screeningmethod, incidence, clinical follow-up procedure, outcomes andcosts. Conclusions regarding each disorder were drawn inrelation to the fulfilment of seven criteria based on the Wilsonand Jungner screening criteria:

(1) the disorder should be clinically and biochemically welldefined;

(2) there should be a known incidence of the disease inpopulations relevant to the United Kingdom;

(3) the disorder should be associated with significantmorbidity or mortality;

(4) effective treatment should be available;(5) there should be a period before the onset of the disease

during which intervention improves outcome;(6) there should be an ethical safe, simple and robust

screening test for the disease;(7) the screening should be cost effective.

Evidence on the last criterion (cost-effectiveness) was availableonly in very few studies and was reviewed separately, as werepapers relating to multi-disease screening technologies. Nopublished controlled intervention studies were identified; theliterature consisted of uncontrolled trials of screening andobservational studies.7 There were no data suitable for meta-analysis. As all the literature was grade III or less,8 inclusioncriteria were content based only.

Results

Disease-based appraisals

The fulfilment of the screening criteria for those IEMs whichare either currently screened for, or for which newborn

SYSTEMATIC REVIEW OF NEONATAL SCREENING 333

StrategyOn-line

Manual

OrganicGrey

MedlineBIDS(Embase, Science CitationIndex to Scientific & TechnicalTextbooksConference proceedingsIndex MedicusCurrent ContentsReferences of referencesTheses, laboratory reports

Literature

Index,Proceedings)

SearchCriteriaKeywords/Textwordsinborn error of metabolism plus one of+ mass screening+ outcome+ incidence+ false positive reactions+ false negative reactions+ costs and cost analysis+ sensitivity and specificity

Citations and Abstracts

1866 references

Review byand by af co-ordinator

subject expert

Exclusion Criteria

Not neonatal screening (unless long-term incidence or outcome data)Not well baby screening (unless long-term incidence or outcome data)Pure laboratory-based studies pre-1980Methodology unsuitable for mass population screening

fInclusion Criteria

Inborn errors of metabolism + defined screening test + data on at least one of• population incidence• effectiveness of screening• health outcomes with or without screening• screening and/or treatment costs• defined screening technologies suitable for blood samples

407 references

Figure 1 Methods of identification and selection of papers.


screening has been suggested in the literature, are listed in Tables1-4. Ideally, for a disease to be identified as suitable for inclusionin a neonatal screening programme, all the criteria should befulfilled. (Because of the lack of cost-effectiveness data for anydisorders other than PKU this criterion is omitted from the tables.)

Table 1 shows that PKU fulfils all the criteria and maytherefore be taken as the standard against which to compareother disorders. Non-PKU aminoacidopathies, because of theiruncertain incidence and a lack of evidence of effectivetreatment, do not fulfil all the criteria. Among the disordersof fatty acid ^-oxidation (Table 2) medium-chain acyl CoAdehydrogenase (MCAD) deficiency shows greatest fulfilmentof the criteria, because of its high incidence and the potentiallarge health benefits from treatment, although the naturalhistory of this disorder is not yet fully known and it has beensuggested that many individuals may remain asymptomatic.9

Other disorders of fatty acid /3-oxidation fail significantly tofulfil the screening criteria. Screening for MCAD deficiency isdependent upon the use of tandem MS. None of the 'common'organic acid disorders (Table 3) show even moderately goodfulfilment of the screening criteria, many presenting acutely inthe newborn period and most showing relatively poor long-termoutcome despite improvements in therapy. There is additionallyno evidence to suggest that early detection and treatmentimprove outcome. Glutaric aciduria type 1 (GA1) may be anexception, as this disorder has a significant asymptomaticperiod and is associated with severe neurological sequelae thatmay be totally preventable by simple treatment. Similar factorsrelate to biotinidase deficiency (Table 4) despite the relativelylow incidence of the complete deficiency.10 In both cases (GA1and biotinidase) therapy is simple, effective and cheap.Screening for GA1 is dependent upon the use of tandem MSwhereas screening for biotinidase deficiency utilizes a simpleassay of enzyme activity.

Congenital adrenal hyperplasia (CAH) caused by deficiencyof the enzyme 21-hydroxylase (Table 4) would completelyfulfil the screening criteria except that for most females there isessentially no asymptomatic period; these infants should beidentified clinically at birth or at the post-natal examination.However, there is still benefit from early detection and diagnosisfrom newborn screening by avoiding diagnostic uncertainty andpreventing gender mis-assignment. Newborn screening pro-grammes for CAH have recently been introduced in bothFrance" and Sweden.12

Previous justification for neonatal screening programmes forgalactosaemia (Table 4) has been based on prevention ofmorbidity; however, published evidence suggests that, despiteearly treatment, long-term outcome is poor, with a continuingdegree of neurological handicap.13 Screening for heterozygousfamilial hypercholesterolaemia (Table 4) has been proposed toallow treatment for this condition to be started before thedevelopment of symptomatic disease. However, measurementof total cholesterol or apolipoproteins in the neonatal period ispoorly predictive of later values and no proven benefits have yet

been shown to arise from treatment of children very early inlife. Similarly, screening for Wilson disease (Table 4) by themeasurement of caeruloplasmin, the copper transportingprotein of plasma, has been proposed but its concentration inneonates with Wilson disease is not yet known. Thus newbornscreening programmes are not currently recommended for anyof these three IEMs.

Review of screening technologies

Several different techniques including bacterial inhibitionassays (Guthrie), chromatography and fluorimetry are currentlyused to screen for PKU. Dried blood spots are generally usedbut some centres use liquid blood samples. A laboratoryscreening for more than one disorder will normally use morethan one technique and the methods used vary from manual topartly automated. Full automation of neonatal screening,whatever technique is used, will require the development ofan automated punch capable of assessing the quality andposition of blood spots on the filter paper.

The technology already exists which could lead to thedevelopment of a fully automated system capable of performingseveral tests. By the use of different labels, time-resolvedfluorescence, currently used for congenital hypothyroidismscreening, can measure up to four analytes simultaneously. Thisnumber of tests can be further widened by the use of a detectorwith photometric and luminometric capabilities in addition tofluorimetry; such an instrument can combine immunoassaywith enzymatic and chemical methods. Such machines will beexpensive (more than £100000) and would only be likely to becost effective in large-scale screening laboratories.

Molecular (DNA) techniques appear generally unsuitablefor universal neonatal screening for IEMs, as these disordersare almost all caused by a very large number of differentindividual mutations, sometimes specific to families, andgenotype-phenotype correlations are not clear. The ideal roleof molecular testing in newborn screening may be forconfirmation in cases identified from more conventionalmethods, especially where the latter are less specific (e.g. forcystic fibrosis).14

Tandem MS has the potential for simultaneous multi-diseasescreening, including PKU, using a single analytical techniqueand would be complementary to immunoassay-based methodsrequired for congenital hypothyroidism, cystic fibrosis andCAH screening.15 Evidence on the current status and future ofpotential neonatal screening by tandem MS was obtained bothfrom the literature review and from visits to four laboratories,two in the United Kingdom and two in the United States,currently utilizing or developing the technique. The procedureused was similar in all the laboratories and all use a punchedsample from the dried blood spot card. The technology has beendemonstrated to be robust (i.e. sensitive and specific) andsuitable for the reliable detection of PKU and certain otherIEMs.16 However, concrete data from prospective newborn

Table 1 Fulfilment of screening criteria for selected amino acid disorders

Amino aciddisorder

Clinically andbiochemicallywell-defineddisorder

Known incidencein populationsrelevant to UK

Associated withsignificant morbidityor mortality

Effective treatmentavailable

Period before onsetduring whichintervention improvesoutcome

Ethical, safe, simpleand robust screeningtest

PKU

Maple syrup urinedisease (MSUD)

Homocystinuria

Tyrosinaemia type 1

• Clinical andbiochemicalspectrum of diseasewell known31

• Clinical andbiochemicalphenotypes welldescribed33

• 1:12 000 (averageUK figure)1

? Average 1:185000world-wide;33

0:467448 screenedin Scotland34

• Natural historynow welldelineated37

• Spectrum ofclinical presentationswell characterized38

together withassociated biochemicalphenotype

? Average 1:291000world-wide37

Estimated 1 234 000in UK, but significantregional variation

?1:10500039

• Severe neurologicaldamage impairingcognitive development;early death

• Developmental delay,seizures andencephalopathy.Presentation in neonatalperiod usually fatalwithout treatment andlater in life death mayoccur during episodesof metabolicdevelopment

• Neurologicaldysfunction,thrombo-embolicevents, impairedvision and ectopialentis

• Liver failure, renaltubular dysfunction,neuropathic crisesand hepatocellularcarcinoma during andafter second decadeof life

• Dietary therapy3

? Dietary therapy,but does notcompletely preventmetabolic crises35

or preventdevelopmentaldelay

? Pyridoxine anddietary therapy withmethionine restrictionand L-cysteinesupplementation

• Early diagnosis andtreatment reduceincidence of neurologicalhandicap from 80-90%to 6-8%32

? Neonatal onset usuallyat 4-7 days (possible beforeresult of screeningavailable)

• Early diagnosis andtreatment improvesIQ by some 35points.37

? Dietary therapyeffective in short term,but alternativetreatments eventuallyneeded, e.g livertransplant. Long-termefficacy of NTBCtreatment still underinvestigation

? Acute onset withinweeks of birth or delayedinto infancy and childhood

• Dried blood spots testedwith bacterial inhibitiontest, chromatography orfluorometry. Tandem MS16

can also be used

• Dried blood spots testedwith bacterial inhibitiontest.34 Tandem MS36 canalso be used

•Dried blood spots forassociated hyper-methioninemia; bacterialinhibition assay40 orchromatography unreliablebut tandem MS muchmore robust41

• Dried blood spots byfluorometry.42

Also tandem MS16

inH

JO

m

m

OTImO

nmmzzo

• Criteria fulfilled.? More data required.X Criteria not fulfilled.

Table 2 Fulfilment of screening criteria for selected disorders of fatty acid /3-oxidation

Fatty aciddisorder







Medium-chain acylCoA dehydrogenase(MCAD) deficiency

Mitochondrial long-chain fatty acidoxidation disorders

Multiple acyl CoAdehydrogenasedeficiency (MACDD)

• Genetically andclinically fairly wellcharacterized.Incidence andlong-term outcomeunknown forasymptomaticindividuals

X Clinical andbiochemical featuresvariable and diverse.Outcome poor tounknown

X Presentation highlyvariable and naturalhistory unknown

71:6500 to 1:20 000(estimated)43

X Unknown butappears rare

X Unknown butappears rare

• Life-threateninghypoglycaemicencephalopathy fromneonatal period. Acuterhabdomyolosis inadults now alsorecognized

? Hypoglycaemia,encephalopathy, earlydeath in some cases.Cardiomyopathy in olderpatients; others lesssevere disease

? Presentation rangesfrom death in infancywith acidosis,hypoglycaemia,hyperammonaemia,convulsions andneurological symptomsto much milder diseasein later childhood/adultlife

• Regular high-carbohydrate intake,emergency dietaryregimen for illness,L-carnitine therapy

? Treatment of mildercases may beeffective but outcomein severe cases ispoor

? Milder cases mayrespond to anincreased carbohydrateand reduced fat andprotein intake,riboflavm andL-carnitine. Severeinfantile diseaseoften refractory totherapy

• Although presentation innewborn period nowdescribed, large majorityare asymptomatic duringthe period of newbornscreening

X Not known for allconditions andpresentations

X Severe neonataldisease often presentsbefore screening resultwould be available.Associated dysmorphicfeatures and polycystickidneys in some casesimply intrauterinedamage

• Dried blood spots;determination ofoctanoylcarnitine usingtandem MS43

? Dried blood spots;determination of long-chainacylcarnitmes (C,,»-C18)using tandem MS butthese are currentlydifficult to detect and notall cases may be identified

• Dried blood spots;determination of relevantacylcarnitines usingtandem MS possible butnot yet proven

oc50Z>rocroxmrX2manzrn

MJ>S References throughout.• Criteria fulfilled.? More data required.X Criteria not fulfilled.

Table 3 Fulfilment of screening criteria for selected organic acid disorders

Organic aciddisorder

Methylmalonicaciduria

Propionic acidaemia

Isovaleric acidaemia

Glutaric aciduriatype 1

3-hydroxy-3-methyl-glutaric aciduria


? Generally wellcharacterized anddefined althoughlong-term outcomestill unclear

? Reasonably wellcharacterized anddefined. Long-termoutcome in 'milder'cases unknown

• Well characterized

? Becoming bettercharacterized butcomplete naturalhistory unknown

? Long-term outcomewith therapy unknown


7 Not known withcertainty but probably<1:80000

7 Probably<1:100000

?Very low in UK,probably <1:200000

? Unclear but may beas high as 1:50000

? Not known,probably < 1.80 000


• Life-threateningepisodes of acuteillness with longer-term problems ofneurological handicapand renal damage

• Life-threateningepisodes of acuteillness with longer-term problems ofneurological handicap

• Presents withacidosis, vomiting,tremors, coma anddeath in the neonatalperiod and beyond

• Severe cripplingchoreoathetoid cerebralpalsy of postnatal onset

• Episodic hypokekotichypoglycaemia causingcoma and death andpossible long-termneurological sequelaeinto at least teenageyears


7 Improving with diet,antibiotics, L-carnitine,somatotrophin, organtransplantation, andB12 in vitaminresponsive cases

? Improving with diet,antibiotics, L-carnitine,somatotrophin, livertransplantation, B12

in vitamin responsivecases

• Diet plus glycineand L-carnitinetherapy

? Current opinionsuggests L-carnitinetherapy pre-symptomatically mayprevent onset ofneurological damage

• Moderate dietarymodification,emergency regimenfor illness, andL-carnitine therapy.Need for lifelongtherapy unknown


? Many cases present inneonatal period, otherslater in infancy andchildhood. No evidenceearly interventionimproves outcome

7 Many cases present inneonatal period. Evidenceof improved outcomefollowing earlyintervention notavailable

7 Many cases presentacutely in neonatal period.No evidence earlyintervention improvesoutcome but will preventacute episodes

• All patients seem tohave an asymptomaticperiod of a month ormore

7 Generally there isasymptomatic periodthough neonatalpresentation isrecognized


• Dried blood spots;determination ofpropionylcarnitines usingtandem MS but levels maybe very low in some cases

• Dried blood spots;determination ofpropionylcarnitines usingtandem MS

• Dried blood spots;determination ofisovalerylcarnitines usingtandem MS

• Dried blood spots;determination ofglutarylcarnitines bytandem MS

• Dried blood spots;determination of 3-methylglutarylcarnitines usingtandem MS

«<

H

m2>(-}

J0m<

o

mo2>

>r-on

npatnrnZZO

29.44.46.47 R e f e r e n c e s throughout.• Criteria fulfilled.7 More data required.X Criteria not fulfilled.

Table 4 Fulfilment of screening criteria for other metabolic disorders

Metabolicdisorder







Biotinidase deficiency •Causes multiple ?Estimated in UKcarboxylase deficiency, about I I I IOOOO 4 8

Long-term outcomeof therapy to be fullyevaluated

Congenital adrenalhyperplasia (CAH)

Familialhypercholesterolaemia(FH)

• Well defined.Refers to21-hydroxylasedeficiency

• Well-characterizedcause of hyper-cholesterolaemia

Galactosaemia • Very well-defineddefect

Wilson disease • Well-characterizeddefect in coppertransport

• Estimated 1.6000to1:2100049-50

• Homozygotes1:100000052

Heterozygotes 1:200to 1:1000

1:4400013

• 1:50000 to1:10000055

• Progressiveneurological diseasewhich may be fatalin many cases

• Hyponatraemicdehydration,vinhzation

• Heterozygotes:vascular damage(atherosclerosis)estimated to cause 5%of premature coronaryheart disease (CHD)53

Approx. 50% men withFH develop CHD by age50, 50% women by 60Homozygotes: fatal

• Liver and renaldamage, cataractformation,immunodeficiency,neurological damage

• Liver failureneurological damage

• Oral biotm.Requirements forlife-long therapynot known

• Steroid replacement

? Benefit fromintervention(heterozygotes) duringchildhood yet to bedemonstrated

X Dietary therapyreverseshepatotoxicity andprevents cataractdevelopment but notprogressive neuro-degenerative disorder56

or ovarian failure

• Penicillamine.Pre-symptomatictreatment with zincmay be possible57

• Presentation generallyin early infancy andchildhood

? Clinically present at 3days to 8 weeks, mostcommonly at 3 weeks

? Coronary heartdisease onlyangiographicallydetectable in youngadult life(heterozygotes)

? Present typically insecond week of life

• Most cases presentin late childhood,adolescence and earlyadulthood58-59

• Dried blood spots;colorimetric assay ofenzyme activity.51 Need todistinguish subjects withpartial deficiency

• Determination of 17a-hydroxy progesteronefrom dried blood spots11

XTotal or LDLcholesterol orapolipoprotem B can bemeasured from bloodspots54 but are poorlypredictive of latervalues

? Colorimetric enzymeassay on dried bloodspots.60 False positiverate 0.03%, false negativerate 0-7%. Beutlerfluorescent assay61

? Enzymatic assay ofcaeruloplasmin62 orELISA method63 but notyet validated for use innewborns

Oa2

o-aCCO

rXm>rx2mDnzm

• Criteria fulfilled.? More data required.X Criteria not fulfilled.

Table 5 Collective fulfilment of criteria and recommendations for newborn screening

Inborn error of metabolism Fulfilment of screening criteria and recommendation for newborn screening

PKU

Maple syrup urine disease (MSUD)

Homocystinuria

Tyrosinaemia type 1

Medium-chain acyl CoA dehydrogenase(MCAD) deficiencyMitochondrial long-chain fatty acid oxidationdisorders

Multiple acyl CoA dehydrogenase deficiency(MACDD)

Methylmalonic aciduria 1Propionic acidaemia J

Isovaleric acidaemia

Glutaric aciduria type 1 *

3-hydroxy-3-methyl-glutanc aciduria

Biotinidase deficiency*

Congenital adrenal hyperplasia (CAH)*

Familial hypercholesterolaemia (FH)

Galactosaemia

Wilson disease

(1) Fulfils all screening criteria (including cost-effectiveness)

(3) Low incidence, neonatal onset often pre-dates availability of screening result, long-term outcome of treatment variable.

(3) Low incidence and effectiveness of treatment over a lifetime unproven

(3) Low incidence, no evidence of improved outcome from early, pre-symptomatic treatment

(2) High frequency, morbidity and mortality largely preventable by simple, cheap and effective treatment, simple screening testrequires tandem MS. Screening research programme required to assess exact incidence, long-term outcomes and cost-effectiveness(4) Unknown natural history, low incidence, ineffective treatments in many cases and possible ambiguous screening results withpossible false negatives using tandem MS

(4) Widely variable clinical presentation, some cases with intrauterine damage, dysmorphology, acute neonatal presentation,ineffective treatment, low unknown incidence and unproven screening test (although possible with tandem MS)

(3) Highly variable outcome, many cases present in newborns before screening result available, treatment efficacy variable. At best,potential candidates for newborn screening. Incidence has to be considered collectively with other similar disorders detectable usingtandem MS. Tandem MS does not distinguish these two organic acid disorders(3) Very low incidence offset by collective incidence of other disorders; early treatment may prevent acute episodes and preventneurological damage Potential candidate for newborn screening, requires tandem MS

(3) Potentially preventable severe neurological and disabling disorder, thus a candidate for newborn screening despite probable lowincidence. Assessment of screening research programme required to ascertain exact incidence in UK population and evaluateprevention of neurological disease in screened and treated individuals. Requires tandem MS

(3) Fairly low incidence but simple treatment with prevention of neurological damage makes this a potential candidate for newbornscreening. Incidence must be considered collectively as above; requires tandem MS

(2) Preventable severe neurological disorder, thus a candidate for newborn screening despite estimated low incidence. Screeningresearch programme required to assess exact incidence in UK, long-term outcomes and cost-effectiveness(2) To fully meet all criteria may need screening at an earlier age (e.g. day 3) to allow availability of results by 10-12 days of age(4) No suitable validated neonatal screening test Benefit of intervention in neonatal period not determined

(4) Therapy does not prevent neurological or ovarian toxicity; long-term outcome does not benefit from early intervention throughneonatal screening(3) Low incidence No suitable validated neonatal screening test

Hm

>H

om<m

ozmoz

n73tnmzzo

(1) Evidence that all criteria are met.(2) Evidence that all criteria except known natural history of disorder are met.(3) Evidence that most criteria are met, but lack of evidence for some criteria.(4) Evidence that some criteria are not met."Trials of neonatal screening recommended.


screening using tandem MS is lacking or unsubstantiated, andfurther research through large-scale, well-conducted and exter-nally assessed trials is required before this technology can beuniversally introduced into newborn screening programmes.

Economic evidence

There was insufficient evidence in the literature to assess theeconomic value of screening for any disorder other than PKU.Twelve economic evaluations of neonatal screening for inbornerrors of metabolism were identified.17"28 The literaturecovered a range of disorders, but only PKU was included inmore than one study. There were eight cost-benefit studies ofneonatal screening for PKU,1 7"2 1 '2 3 '2 5 '2 7 all of which concludedthat PKU screening is worth while in monetary terms alone. Notall of the papers included sample collection costs, but wherethey did these were outweighed by the net benefits of screening.This suggests that PKU screening alone justifies the collectionof blood samples from neonates.

Discussion

The striking finding of this review was the lack of robustevaluation data. The absence of controlled trials meant reliancehad to be put on observational studies. As it is difficult toquantify the rigour of this type of research, all being grade HI orless,8 all studies were included if they fulfilled the contentcriteria. Agreement on the Wilson and Jungner6 criteria wasthen achieved by paired reviewers and eventually with thewhole group. As there remains a subjective element because ofthe types of study used, underlying references have beenprovided so that readers can refer to the specific papers.Nevertheless, some evidence has been obtained for theappropriateness or otherwise of screening for a wide range ofIEMs. There is clear evidence that all of the criteria are met forPKU screening, which provides a positive net benefit to theindividual and to society. In addition, PKU screening by itselfjustifies the infrastructure of sample collection and testing ofnewborn blood spots.29

Table 5 summarizes the results and recommendations foreach disorder. Only four disorders showed adequate fulfilmentof the screening criteria to be considered for inclusion inexpanded newborn screening programmes. These were biotini-dase deficiency, CAH, MCAD deficiency and, possibly, GALFor these disorders structured, co-ordinated and continuingevaluation, including economic analysis, will be necessary toprovide evidence which would justify the long-term continua-tion of these programmes. A number of other disorders,including some of the more common disorders of organic acidmetabolism, require further basic research before widespreadtrials of neonatal screening could be advocated.

Concern has also been expressed about the lack ofconsideration of the infrastructure required to deal with theinfants identified by newborn IEM screening. In many areas of

the United Kingdom, for example, there appears to be relativelypoor liaison between the screening laboratories, midwives andother health-care personnel, and no provision is made for theco-ordinated follow-up or management of identified patients.30

Thus there is a need for a better infrastructure for notificationand continued care, including parental counselling, of patientswith IEMs identified through newborn screening. Few areasprovide adequate information to parents about newbornscreening or the IEMs screened for, nor do they currentlyrequire informed consent from the parents;5 issues which willneed to be addressed if expanded screening programmes areintroduced.

Newborn screening for MCAD and GA1 depends upon theuse of tandem MS technology. Pressure to develop andintroduce this technology is particularly acute in countriessuch as the United States where blood spots are taken within thefirst 24-48 hours after birth. However, the utility andapplication of tandem MS for prospective newborn screeninghas been demonstrated in only one centre (in the United States),with only a limited number of newborns screened.29 Thetechnology therefore requires further assessment throughprimary research before it can be considered for universalintroduction. The size of screening laboratories varies widely;for tandem MS technologies to be cost-effective eachlaboratory would need to perform a minimum number of testsper year, necessitating rationalization of current screeninglaboratories. There is a need for co-ordinated national policiesfor newborn screening that ensures the appropriate dissemina-tion of new screening technologies and the expansion of theroutine screening programme where there is good evidence ofthe efficacy of early detection and treatment. Any such changesmust be accompanied by careful evaluation to ensure theireffectiveness and cost-effectiveness in practice.

Acknowledgement

This work was funded by the NHS Health TechnologyAssessment Programme.

References

1 Smith I, Cook B, Beasley M. Review of neonatal screeningprogramme for phenylketonuria. BrMedJ 1991; 303: 333-335.

2 Novello AC. Inherited metabolic diseases: collaborating forthe health of all children. Biochem Med Metab Biol 1993;49: 277-284.

3 Streetly A, Grant C, Pollitt RJ, Addison GM. Survey ofscope of neonatal screening in the United Kingdom. Br MedJ 1995; 311: 726.

4 Tappin DM, Girdwood RWA, Kennedy R, et al. An audit ofinfants 'unscreened' by the neonatal screening system inScotland during 1991. Screening 1995; 3: 201-207.


5 Clayton EW. Issues in state newborn screening programs.Pediatrics 1992; 90: 641-646.

6 Wilson JM, Jungner G. Principles and practice of screeningfor disease. Public Health Papers, 34. Geneva: World HealthOrganization, 1968: 26-39.

7 Woolf SH, Battista RN, Anderson GM, Logan AG, Wang E.Assessing the clinical effectiveness of preventive maneu-vers: analytic principles and systematic methods inreviewing evidence and developing clinical practicerecommendations. A report by the Canadian Task Force onthe Periodic Health Examination. J Clin Epidemiol 1990;43: 891-905.

8 Canadian Task Force on the Periodic Health Examination.The Periodic Health Examination. Can Med Assoc J 1979;121: 1193-1254.

9 Seddon HR, Green A, Gray RG, Leonard JV, Pollitt RJ.Regional variations in medium-chain acyl-CoA dehydro-genase deficiency. Lancet 1995; 345: 135-136.

10 WarnerRogers J, Waisbren SE, Levy HL. Cognitivefunction in early treated biotinidase deficiency: follow-up ofchildren detected by newborn screening. Screening 1995; 4:125-130.

11 Pang S, Clark A. Congenital adrenal hyperplasia due to 21-hydroxylase deficiency: newborn screening and its rela-tionship to the diagnosis and treatment of the disorder.Screening 1993; 2: 105-139.

12 Nordenstrom A, Thilen A, Hagenfeldt L, et al. Benefits ofscreening for congenital adrenal hyperplasia in Sweden. In:Levy HL, Hermos RJ, Grady GF, eds. Third InternationalMeeting of the Society for Neonatal Screening. Boston, MA:Third International Society for Neonatal Screening, 1996:211-212.

13 Honeyman MM, Green A, Holton JB, Leonard JV.Galactosaemia: results of the British PaediatricSurveillance Unit Study, 1988-90. Arch Dis Child 1993;69: 339-341.

14 Makowski GS, Aslanzadeh J, Hopfer SM. In situ PCRamplification of Guthrie card DNA to detect cystic fibrosismutations. Clin Chem 1995; 41: 477-479.

15 Ostrup J, Jones M, Martinmaki A. A concept for fullautomation of immunoassays performed in newbornscreening laboratories. In: Levy HL, Hermos RJ, Grady GF,eds. Third Meeting of the International Society for NeonatalScreening. Boston, MA: Third International Society forNeonatal Screening, 1996: 171-172.

16 Chace DH, Millington DS, Terada N, et al. Rapid diagnosisof phenylketonuria by quantitative analysis for phenylala-nine and tyrosine in neonatal blood spots by tandem massspectrometry. Clin Chem 1993; 39: 66-71.

17 Aim J, Larsson A, Rosenqvist U. Health economic analysisof the Swedish neonatal metabolic screening programme.

A method of optimizing routines. Med Decision Making1982; 2: 33-45.

18 Anonymous. Massachusetts Department of Public Health.Cost-benefit analysis of newborn screening for metabolicdisorders. N Engl J Med 1974; 291: 1414-1416.

19 Barden HS, Kessel R, Schuett VE. The costs and benefits ofscreening for PKU in Wisconsin. Social Biol 1984; 31: 1-17.

20 Bush JW, Chen MM, Patrick DL. Health status index in costeffectiveness: analysis of PKU program. In: Berg RL, ed.Health Status Indexes. New York: Hospital and EducationalTrust, 1973: 172-209.

21 Cockburn F, Clark BJ, Byrne A, et al. Maternal phenyl-ketonuria - diet dangers and dilemmas. Int Pediatr 1992;7(1): 67-74.

22 Dagenais DL, Courville L, Dagenais MG. A cost-benefitanalysis of the Quebec Network of Genetic Medicine. SocialSciMed 1985; 20: 601-607.

23 Dhondt JL, Farriaux JP, Sailly JC, Lebrun T. Economicevaluation of cost-benefit ratio of neonatal screeningprocedure for phenylketonuria and hypothyroidism. JInherited Metab Dis 1991; 14: 633-639.

24 Dunkel G, Scriver CR, Clow CL, et al. Prospectiveascertainment of complete and partial serum biotinidasedeficiency in the newborn. J Inherited Metab Dis 1989; 12:131-138.

25 Hisashige A. Health economic analysis of the neonatalscreening program in Japan. Int J Technol Assessment HlthCare 1994; 10: 382-391.

26 Holtzman NA. Pitfalls of newborn screening (with specialattention to hypothyroidism): when will we ever learn?Birth Defects: Original Article Series 1983; 19: 111-120.

27 Komrower GM, Sardharwalla IB, Fowler B, Bridge C. TheManchester regional screening programme: a 10-yearexercise in patient and family care. BrMedJ 1979; 2: 635-638.

28 Sprinkle RH, Hynes DM, Konrad TR. Is universal neonatalhemoglobinopathy screening cost-effectiveness? ArchPediatr Adolescent Med 1994; 148: 461-469.

29 Thomason MJ, Lord J, Bain MD, et al. Newborn screeningfor inborn errors of metabolism: a systematic review. HealthTechnol Assess 1997; 1(11).

30 Cappuccio FP, Hickman M, Barker M. Neonatal screening.Performance is hard to monitor. Br Med J 1996; 312: 182.

31 Scriver CR, Kaufman S, Eisensmith RC, Woo SLC. Thehyperphenylalaninemias. In: Scriber CR, Beaudet AL,Sly WS, Valle D, eds. The Metabolic and MolecularBases of Inherited Disease, 7th edn. New York: McGraw-Hill, 1995: 1015-1075.

32 Smith I, Brenton DP. Hyperphenylalaninaemias. In:


Fernandes J, Saudubray J-M, van der Berghe G, eds. InbornMetabolic Diseases. Berlin: Springer-Verlag, 1995: 147-160. •

33 Chuang DT, Shih VE. Disorders of branched chain aminoacid and keto acid metabolism. In: ScriverCR, Beaudet AL,Sly WS, Valle D, eds. The Molecular and Metabolic Basesof Inherited Disease, New York: McGraw-Hill, 1995:1239-1277.

34 Naylor EW, Guthrie R. Newborn screening for maple syrupurine disease (branched-chain ketoaciduria). Pediatrics1978; 61: 262-266.

35 Treacy E, Clow CL, Reade TR, et al. Maple syrup urinedisease: interrelations between branched-chain amino-, oxo-and hydroxyacids; implications for treatment; associationswith CNS dysmyelination. J Inherited Metab Dis 1992; 15:121-135.

36 Chace DH, Hillman SL, Millington DS, et al. Rapiddiagnosis of maple syrup urine disease in blood spots fromnewborns by tandem mass spectroscopy. Clin Chem 1995;41: 62-68.

37 Mudd SH, Skovby F, Levy HL, et al. The natural historyof homocystinuria due to cystathionine beta-synthasedeficiency. Am J Human Genet 1985; 37: 1-31.

38 Mitchell GA, Lambert M, Tanguay RM. Hypertyrosinemia.In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. TheMetabolic and Molecular Bases of Inherited Disease, 7thedn. New York: McGraw-Hill, 1995: 1077-1106.

39 Hutchesson ACJ, Hall SK, Preece MA, Green A. Screeningfor tyrosinemia type 1. Arch Dis Child 1996; 74: F191-F194.

40 Whiteman PD, Clayton BE, Ersser RS, Lilly P, Seakins JW.Changing incidence of neonatal hypermethioninaemia:implications for the detection of homocystinuria. Arch DisChild 1979; 54: 593-598.

41 Chace DH, Hillman SL, Millington DS, et al. Rapiddiagnosis of homocystinuria and other hypermethionine-mias from newborns' blood spots by tandem mass spectro-scopy. Clin Chem 1996; 42: 349-355.

42 Laberge C, Grenier A, Valet JP, Morissette J. Fumaryl-acetoacetase measurement as a mass-screening procedurefor hereditary tyrosinemia type I. Am J Human Genet 1990;47: 325-328.

43 Ziadeh R, Hoffman EP, Finegold DN, et al. Medium chainacyl-CoA dehydrogenase deficiency in Pennsylvania: neo-natal screening shows high incidence and unexpectedmutation frequencies. Pediatr Res 1995; 37: 675-678.

44 Chalmers RA. Disorders of organic acid metabolism. In:Holton JB, ed. The Inherited Metabolic Diseases. Edin-burgh: Churchill Livingstone, 1994: 115-204.

45 Roe CR, Coates PM. Mitochondrial fatty acid oxidation

disorders. In: Scriver CR, Beaudet AL, Sly WS, Valle D,eds. The Metabolic and Molecular Bases of InheritedDisease, 7th edn. New York: McGraw-Hill, 1995: 1501-1533.

46 Sweetman L, Williams JC. Branched Chain OrganicAcidurias. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds.The Metabolic And Molecular Bases of Inherited Disease, 7thedn. New York: McGraw-Hill, 1995: 1387-1422.

47 Fenton WA, Rosenberg LE. Disorders of propionate andmethylmalonate metabolism. In: Scriver CS, Beaudet AL,Sly WS, Valle D, eds. The Metabolic and Molecular Basesof Inherited Disease, 7th ed. New York: McGraw-Hill,1995: 1423-1449.

48 Wolf B. Worldwide survey of neonatal screening forbiotinidase deficiency. J Inherited Metab Dis 1991; 14:923-927.

49 Rudd BT. Prevalence of adrenal 21-hydroxylase deficiencyin neonates born in the West Midlands: a retrospectivestudy. J Inherited Metab Dis 1986; 9(Suppl 1): 155-156.

50 Wallace AM. Review of CAH screening programmes andthe Scottish experience. J Inherited Metab Dis 1986;9(Suppl 1): 135-141.

51 Pettit DA, Amador PS, Wolf B. The quantitation of biotinidaseactivity in dried blood spots using microtiter transfer plates:identification of biotinidase-deficient and heterozygousindividuals. Analyt Biochem 1989; 179: 371-374.

52 Goldstein JL, Hobbs HH, Brown MS. Familial hyper-cholesterolemia. In: Scriver CR, Beaudet AL, Sly WS, ValleD, eds. The Metabolic and Molecular Bases of InheritedDisease, 7th edn. New York: McGraw-Hill, 1995: 1981-2030.

53 Gillman MW. Screening for familial hypercholesterolemiain childhood. Am J Dis Child 1993; 147: 393-396.

54 Asami T. Screening for hypercholesterolaemia on bloodspotted filterpaper. Lancet 1983; 2: 229-230.

55 Reilly M, Daly L, Hutchinson M. An epidemiological studyof Wilson's disease in the Republic of Ireland. J NeurolNeurosurg Psychiat 1993; 56: 298-300.

56 Segal S. Galactosemia unsolved. Eur J Pediatr 1995; 154:S97-S102.

57 Brewer GJ, Dick RD, Yuzbasiyan-Gurkan V, Johnson V,Wang Y. Treatment of Wilson's disease with zinc. XIII:Therapy with zinc in presymptomatic patients from the timeof diagnosis. J Lab Clin Med 1994; 123: 849-858.

58 Yarze JC, Martin P, Munoz SJ, Friedman LS. Wilson'sdisease: current status. Am J Med 1992; 92: 643-654.

59 Danks DM. Disorders of copper transport. In: Scriver CR,Beaudet AL, Sly WS, Valle D, eds. The Metabolic andMolecular Bases of Inherited Disease, 7th edn. New York:McGraw-Hill, 1995: 2211-2235.


60 Schickling H, Schweitzer S, Shin Y, Diepenbrock F, Sander disease and Menkes' kinky-hair disease. Lancet 1977; 2:J. Neonatal screening for galactose metabolic defects. In: 1140.Farriaux J-P, Dhondt J-L, eds. New Horizons in Neonatal 6 3 F u j i o k a Y , Aoki T, Shimizu N, et al. A new screeningScreening. Amsterdam: Elsevier, 1994: 185-189. method for Wilson's disease by measuring blood caerulo-

61 Thibodeau DL, Andrews W, Meyer J, Mitchell P, Wolf B. plasmin level. In: Farriaux J-P, Dhondt J-L, eds. NewComparison of the effects of season and prematurity on the Horizons in Neonatal Screening. Amsterdam: Elsevier,enzymatic newborn screening tests for galactosemia and 1994: 285-288.biotipidase deficiency. Screening 1993; 2: 19-27.

62 Aoki T, Nakahashi M. New screening method for Wilson's Accepted on 9 March 1998

a systematic review of evidence for the appropriateness of neonatal screening ... › d164 ›...

Documents