Arch. Dis. Childh., 1967, 42, 416.

Changes in Blood Gas Tensions FollowingAdministration of Amine Buffer THAM to Infants with

Respiratory Distress SyndromeJ. M. GUPTA*, G. W. DAHLENBURGt, and J. A. DAVIS

From the Nuffield Neonatal Research Unit, Institute of Child Health, Hammersmith Hospital,Du Cane Road, London W.12

A number of workers have shown that theprognosis in respiratory distress syndrome (RDS)is more closely related to the arterial oxygentension than to acid-base disturbance, and dependsto a large extent on the degree of shunting of venousblood past the gas-exchanging portions of lung, i.e.on the extent to which the arterial Po2 responds toan increase in the concentration of oxygen breathed(Boston, Geller, and Smith, 1966; Gupta, 1966;Moss, Emmanouilides, Retorri, Higashino, andAdams, 1965). These findings are to some extentin conflict with claims that the prognosis is improvedby correction of the secondary acid-base distur-bance by infusion of glucose/bicarbonate solutions(Usher, 1959; Hutchison, Kerr, Douglas, Inall, andCrosbie, 1964). It occurred to us that these twopoints of view might be reconciled if it could beshown that correction of pH leads to improvementin oxygenation and diminution of shunting, aconsequence that might be predicted in the new-born from the dependence of pulmonary perfusiononpH and blood gas tensions (Cassin, Dawes, Mott,Ross, and Strang, 1964) as well as on adequatecardiac function. THAM rather than bicarbonatewas used because of its greater alkalinity and itsability to lower the PaCo2 (Nahas, 1959; Holmdahl,Nahas, Hassam, and Verosky, 1961) as well as toraise the pH. It was decided to measure Po2 levelsin arterial blood before and after administration ofTHAM and to see whether any improvement inPaO2 so achieved would be sustained. During thecourse of the study the work of Chu, Clements,Cotton, Klaus, Sweet, Thomas, and Tooley (1965)in Singapore became known to us, and reinforcedour belief that pulmonary hypoperfusion might be a

Received January 3, 1967.* Percy J. Neate Research Fellow (Clothworkers' Company).t Supported by the Frank J. Duval Travelling Fellowship

(Melboume).

major or reversible factor in the pathogenesis ofRDS. The results reported below do, in ouropinion, bear out this view.

Materials and MethodsThe subjects of this investigation were newborn

infants admitted to the neonatal ward of the Hammer-smith Hospital. Most of the infants were born in thehospital but a few were admitted from the district orfrom other hospitals. The infants born at HammersmithHospital were assessed at one minute using a modifica-tion of the Apgar method (Tizard, 1964).

All infants were nursed in incubators. The incubatortemperature varied between 32 and 350 C. but was keptconstant for a given baby at a level as close as possible tothe neutral temperature zone appropriate to its birth-weight (Scopes and Ahmed, 1966). At the time thestudy was made it was the rule to maintain the humidityat nearly 100%. The infants were kept under constantobservation by trained nurses during the period ofinvestigation. The observations consisted of 1/4-hourlyrespirations, 1/2-hourly apex beat, hourly rectal tempera-ture, and general comments on the condition of the baby,i.e. skin colour, muscle tone, response to stimuli,irritability, and character of respiration. The first feed(usually at 6 hours) consisted of water. This wasfollowed by full-strength expressed breast milk 20 ml./kg.divided into 8 feeds, with a daily increment of 20 ml./kg.,so that at 8 days the infants were receiving 160 ml./kg.per day. This regimen was well tolerated.A clinical diagnosis of RDS was made if two of the

following three criteria were fulfilled on assessment atthe age of 4 hours: (1) a respiratory rate of 60 or more ontwo successive observations, or a rise in respiratory rateof more than 20 breaths per minute from 1 to 4 hours ofage; (2) intercostal or subcostal recession; and (3)grunting on expiration.An x-ray film of the chest was obtained as early as

possible and frequent clinical assessment was carried outto confirm the diagnosis and to distinguish RDS fromother causes of dyspnoea, such as intrapartum pneumo-nia, requiring specific therapy. All babies who died hada necropsy.

416

Effects of THAM in Respiratory Distress SyndromeArterial blood samples were collected from the iliac

artery or the abdominal aorta by means of a polyvinylcatheter, size French 5 (Pharmaseal K32 prematureinfant feeding tube) inserted into an umbilical artery.The catheter was securely stitched to the umbilicalstump and the open end was closed with a rubber bung.The cord stump and the surrounding skin were sprayedwith 'polybactrin' solution and covered with a steriledressing.

Umbilical arterial blood samples of 0 4-0 *6 ml. weretaken under anaerobic conditions with a 2 or 5 ml.paraffin-lubricated glass syringe, the dead space of whichhad been filled with heparin (1000 units/ml.). Thesyringe was then sealed with a metal cap. Mostmeasurements of arterial oxygen tensions (Pao2) and pHwere made immediately, the laboratory being adjacent tothe neonatal ward. They were not routinely made induplicate, as this would have involved taking undulylarge samples. Paco2 and standard bicarbonate weredetermined within 1 hour. It was aimed to take bloodsamples at 3-hourly intervals for the first 12 hours and6-hourly intervals thereafter until the catheter wasremoved, but in practice it was found necessary tosample more frequently in severely ill babies and theabove regime was not strictly adhered to for a number ofpractical reasons. After taking the blood sample thecatheter was filled with heparinized saline solution(10 units/ml.) to prevent clotting.

Before collection of a blood sample the baby's rectaltemperature was taken with a clinical thermometer(inserted 3-4 cm.) and the oxygen concentration in theincubator was measured with a Beckman D2 oxygenanalyser. PaO.2 measurements were carried out with aBeckman 02 macro-electrode which was calibrated byusing gas mixtures containing known concentrations ofoxygen. Cord blood was equilibrated with the samegas mixtures in a tonometer and the Po2 was measuredafter equilibration. The average difference in Po2between gas and blood equilibrated with the same gaswas determined and a correction factor was applied (thiswas less than 3%). The gas mixtures were analysed bythe Lloyd-Haldane technique. All the P5o2 valuesmeasured, except those above 200 mm. Hg were conver-ted by using a nomogram (Bradley, Stupfel, andSeveringhaus, 1956) to patient's rectal temperature.For PaO2 values above 200 mm. Hg the temperaturecorrections recommended by Hedley-Whyte and Laver(1964) have been used.

Arterial bloodpH was measured at 370 C. with a directreading pH meter (Electronic Instruments Ltd.)correcting for temperature by adding or subtracting0-0147 for each degree centigrade by which the bodytemperature differed from 370 C. (Rosenthal, 1948). Bymeasuring the change in corrected pH after the samplehad been equilibrated (using a home-made apparatus,Stevens and Lanning, 1964) with 3-50/o and 6-5o% CO2in oxygen, PaCo2 was derived by means of interpolation(Peters, 1923; Astrup, 1956). No significant differencewas observed between the calculated Paco2 values anddirect simultaneous measurements made with a CO2electrode.

THAM was administered as a 3 - 60o solution (0 *3M),the pH of which had been adjusted to approximately 8 8at 370 C. by admixture with its hydrochloride. Dosagewas empirical, the initial dose based on initial pH andbirthweight amounting to approximately 1 ml./kg. foreach pH unit below 7-4. Further doses were givenaccording to the changes in PaO2, pH, and Pco2 achieved.The properties of THAM and the progressive nature ofthe illness are such that it was decided not to base thedosage on the apparent base deficit; however, THAMwas not given unless arterial Pco2 was above 40 mm., pHlower than 7 - 25, and PaO2 indicated a considerable shunt.

Early in the study THAM was given either by theumbilical artery or vein, but in the latter stages entirelyby the umbilical vein because of apparently morebeneficial results. The rate of injection was approxi-mately 1 ml./min., and every dose was followed by a'chaser' of saline to clear the dead space. Blood sampleswere taken immediately before giving THAM and againshortly afterwards, the average interval being 12 minutes(range 4-30 minutes). Careful histological examinationin those babies who subsequently died failed to show anymacro- or microscopic evidence of damage to vessel walls(whether artery or vein) or to the tissues supplied bythem (such as the liver in the case of the umbilical vein,or the leg muscles in the case of the artery).

Surface activity of lung extracts was measured with aWilhelmy balance, in two cases supplemented by staticpressure volume curves.

Blood pressure measurements were made using theMinogograph 42B electromanometer. The transducerwas connected to the catheter in the umbilical arteryafter previously being standardized with a mercurymanometer.

Determination of arterial Po2 at alveolar Po2660 mm. Hg (i.e. breathing 100% oxygen with alveo-lar Pco2 40 mm. Hg). The shunt equation can be usedto calculate the relation between alveolar Po2 (PAo2)and arterial Po2 (PaO2) (Nelson and Reynolds, 1964).Assuming the arteriovenous oxygen difference to be4 ml./100 ml., and oxygen capacity of Hb to be 20 ml.02 per 100 ml. blood, isopleths (shunt lines) wereconstructed (Fig. 1) by calculating the Pao2 for any givenPAO2 at various degrees of shunting. If the calculationsare made by assuming any other arteriovenous oxygendifference or any other oxygen capacity of the Hb orboth, the isopleths retain the same shape but shift inposition. It follows that if the Pao2 is measured at aknown PAO2 (assuming PAco2 = 40 mm.) one can derivefrom the isopleths (irrespective of the oxygen capacity ofthe blood or the arteriovenous oxygen difference) theexpected Pao2, were the infant to be breathing 100%oxygen (i.e. at PAo2 660 mm. Hg where PAco2 is assumedto be 40 mm. Hg). This provides a useful means forcomparison in babies breathing different concentrationsof oxygen, but it should be emphasized that the values soobtained usually differ from the actual Pao2 when breath-ing 100% oxygen and are, therefore, of purely illustrativevalue. This could be because increased oxygenconcentrations in the alveoli may in themselves decrease

417

Gupta, Dahlenburg, and Davis

. .I . I.

20 60 100 140 180 240270 30040 80 120 160 200

Arterial Po2 (mm.Hg)FIG. 1.-Isopleths (shunt lines) showing relation between alveolar and arterial oxygen tension (for derivation, see text).

The numbers refer to percentage shunts.

shunting. The actual PaCo2 was of course taken intoaccount when placing a particular reading on theappropriate isopleth, and therefore the apparentdiminution of shunting after THAM could not have beencaused by the moderate fall in P.co2 produced by thedrug. For comparative purposes PA02 values reportedin this paper are derived values for PAO2 660 mm. Hg.

Results

We have grouped the infants into 'good prognosiscases' and 'bad prognosis cases' in line with the find-ings of Boston et al., and Moss et al. (1965), who havenoted that nearly all deaths due to RDS alone are inbabies whose Pao2 cannot be raised above 100 mm.Hg in 100% oxygen between 3 and 6 hours after

birth. From Table I it can be seen that the deathsin the group where this degree of oxygenation wasachieved were not due directly to respiratorydistress.

Infants with 'good prognosis' RDS. Therewere 33 infants in this group; 23 survived, 10 died.

(a) Surviving infants of 'good prognosis' RDS group.None of these was given any alkali or buffer (THAMor bicarbonate), despite the fact that in all but twoinfants the pH was less than 7 30 and in threeinfants was less than 7 * 20 when first measured. Inall cases the pH rose spontaneously as the infantachieved a satisfactory arterial oxygen tension.

All these infants were given added oxygen, theconcentration of which was adjusted according to

TABLE I'Good Prognosis' Respiratory Distress

Gesta- Birth Post-mortem FindingsInfant tional weght Associated CommentNo. Age (wg) Conditions Immediate Cause State of Lungs(wk.) (g)of Death24 25 820 _ Intraventricular Atelectasis Satisfactory blood gas tensions

haemorrhage25 29 1680 Hydrops (Rh.) Respiratory failure; Hyaline Twin pregnancy; twin I

cardiac failure membranes stillborn26 30 1560 Apnoeic attacks Massive pulmonary Atelectasis Satisfactory blood gas tensions

haemorrhage until apnoeic attacks at 14 hr.27 32 1240 Apnoeic attacks Intraventricular Atelectasis Satisfactory blood gas tensionshaemorrhage28 32 1240 Apnoeic attacks Intraventricular Atelectasis Satisfactory blood gas tensions

haemorrhage29 33 1840 Severe Rhesus Hypoglycaemia Hyaline Satisfactory blood gas tensions

incompatibility membranes30 34 1460 - Intraventricular Hyaline Satisfactory blood gas tensions

haemorrhage membranes31 36 2070 _ Intraventricular Hyaline Satisfactory blood gas tensions

haemorrhage membranes32 37 2380 Intrapartum Cerebral birth * Atelectasis Maintained with IPPR from

asphyxia (cord injury age of 4 hours; repeatedround neck) fits

33 37 2650 Maternal diabetes Air embolism Hyaline Satisfactory blood gas tensions;membranes air embolism when umbilical

venous catheter open

418

z

aEE

0

0

700-650-600-550-500-450-400-350-300-250-200-150I100

360 400 440I. . I I I i

Effects of THAM in Respiratory Distress Syndromearterial oxygen tensions. The initial PaCO2 valueswere (with one exception) above the mean fornormal infants of birthweight less than 2500 g.(unpublished data).

(b) Fatal case of 'good prognosis' RDS group. Theimmediate cause of death and the lung findings atnecropsy are shown in Table I. Nos. 25 and 32were the only infants not to maintain satisfactoryarterial oxygen tensions in the course of their illness.The remainder, who were recovering from respira-tory distress, as judged by the ability to maintainsatisfactory arterial oxygen tensions with decreasingambient oxygen requirements, died as the result ofsecondary factors, as shown in the Table.THAM was given to infants 24, 25, 26, and 32 in

the terminal stages of their illness. There was atransient, but unsustained, rise in the PaO% and pH.Infant No. 29 was given THAM when his conditiondeteriorated following an exchange transfusion atthe age of 1 hour; satisfactory PaO2 was maintainedfrom 8 to 22 hours when he suddenly collapsed anddied. A blood sample taken at this time showed aglucose level of less than 10 mg./100 ml.

In 4 of those infants dying of intraventricularhaemorrhage THAM was not given at any stage; in1 it was given terminally (infant No. 24).

Infants with 'bad prognosis' RDS. Therewere 25 infants in this group; 10 survived, 15 died.Table II shows the relevant details. The age atwhich the iniitial data were obtained was not 3hours in all cases, for as experience was gainedin the use of THAM, infants were treated asearly as possible in the illness; moreover, at thebeginning of the study THAM was given onlyif thepH fell below 7 * 10, but later it was given whenthe PaO2 was less than 100 mm. Hg in 100% oxygen,when the pH was less than 7 * 20, or when the PaCo2was greater than 45. The initial dose of THAM,the total of subsequent doses, and the gas tensionsand pH before and after THAM are also shown.

In a high proportion of these infants there was arapid and substantial rise in the PaO2 followingTHAM. Fig. 2 shows this response. All values inthis and subsequent figures show the calculated Pao2(derived from isopleths) for an alveolar oxygentension of 660 mm. Hg. (It should be pointed outthat as the arterial oxygen tensions rose followingTHAM the ambient oxygen was lowered.)When THAM was given in the first 8 hours of the

illness any rise in arterial oxygen tension that wasobtained was usually sustained. This effect wasmuch more obvious when THAM was given via theumbilical vein, and was associated with dramaticclinical improvement in infant No. 52 and with less

600.

400-a,

300-

-7&200

0.1

a,

Id

C)

Surviving infants.------ Fatal cases

.-,..5,

! =,,.

---

Be-fore THAM

I*

.:- - I.s;

10

.20

30 -

*40.50.60.75

AfterTTHAMFIG. 2.-Severe RDS. Arterial oxygen tension beforeand after giving THAM in 22 cases. The percentage

shunts are shown (see text).

dramatic but still substantial improvement in Nos.37, 44, and 45.There was no response to the initial dose of

THAM in one infant (No. 45), but a substantial andsustained rise in the PAO2 followed a second dose.ThepH in all these infants rose following THAM

and again this rise was sustained in 8 out of the 10.In one of the other two (No. 44) bicarbonate(10 mEq) was given at the age of 61 hours, withimprovement in the pH. In the other infant whofailed to respond to THAM or bicarbonate, it wasapparent at 20 hours that there was a superimposedpneumonia.The Paco2, which was high in 9 of the 10 cases,

fell following THAM in 8 of them (in one after thesecond dose). Cases 52 and 44 illustrate thesepoints.

Case 52. This infant was born by the breech followinga spontaneous labour. The baby was apnoeic at 1minute of age, with a heart rate of 96/min., and was limpand blue. He was intubated and given intermittentpositive pressure respiration (IPPR) and began breathingshortly afterwards. At 2 hours of age he was still limp,was very cyanosed in 30% oxygen, and had a respiratoryrate of 60/min., with marked expiratory grunting andsevere costal recession. The air entry by auscultationwas negligible. Chest x-ray film was compatible withRDS. The relevant gas tensions and pH changes areshown in Fig. 3. There was a sudden and dramaticimprovement in the infant's condition during theintravenous injection of THAM: he became pink, cried,moved all his limbs, and lost his costal recession. This

419

a- I -

420 Gupta, Dahlenburg, and DavisTABLI

'Bad Prognosi

Gesta- rIAge at Age at Initial TotalInfant tional Birthweight Respiration Resuscitation IDaltaAl Dose THAM oToHaM P e ANo. Age (g.) at 1 min Reunsato Data Dose TH H M Before AfterTHAM ~~~~THAM THAMW(wk.) (hr.) (hr.) (ml.) (ml') (mm. Hg) (mm. Hg)34 26 830 Gasping Intubated 2, 4 5 I.V. 1 13 43 43 -320t35 27 950 Gasping Intubated 2 21 3 I.A. 6 47 4236 28 1080 Apnoeic Intubated 11 2 5 I.V. 5 50 19037 29 1100 - Face mask 02 5 6 5 I.V. 10 84 36538 29 1180 Gasping Intubated 10 11) 5 I.V. 5 39 5039 29 1220 Apnoeic Intubated 2 21 3 I.V. 9 32 75 280t

40 30 1360 Gasping Intubated 3 7 3 I.V. 6 30 4041 32 1640 - Face mask 02 8 81 5 I.V. 33 20 10-4042 32 1640 Regular Nil 5) 39 3 l.A. 3 29 2843 33 1670 Gasping Face mask O2 3 31 5 I.V. 25 27 4644 33 1760 Gasping Intubated 2I 3 5 I.V. 5 60 34045 33 1860 Apnoeic Intubated 3 3) 5 I.A. 10 42 78 172t46 33 1700 Gasping Intubated 4 4.) 5 I.A. 15 18 2047 34 1640 - Face mask 02 4 4)-A- 5 I.V. 30 20 6848 35 1960 Gasping Intubated 11 - _ - 59 -49 35 3045 Gasping Face mask 02 4 14 5 I.V. 5 44 7750 35 2030 Gasping Intubated 5 54 3 I.A. 11 16 1751 35 2400 Gasping Face mask 02 4) 12 5 I.V. 27 40 4252 35 2300 Apnoeic Intubated 2 2) 10 I.V. 10 23 46 350t53 36 2000 Gasping Face mask 02 41 5 5 I.V.' 20 57 50 115t54 37 3520 Apnoeic Intubated 8 81 5 I.A. 10 53 9355 37 2240 Regular; Intubated 41 6 5 I.A. 15 100 93

apnoeic -56 38 3000 Apnoeic Intubated 6 i 5 I.V. 5 32 -57 40 3760 Apnoeic Intubated 6 71 5 I.V. 5 60 80 - 18058 ?44 3420 Gasping Intubated 8 - - - 65

* Average time after THAM 12 minutes. t Response to further doses of THAM. IVH intraventricular haemorrhage.

600 - Ca se 52 0-O500 - O/400 0

EE 300 -0C10o E 200 -cE

E 72C n pH A A

O100 .0 p

7-0

0 _ _ _ _ _ _

_6_82 3 4 5 8 12 16 20 24 28Age in hours

FIG. 3.-Case 52. Arterial oxygen tension, pH, and Pco., before and after THAM (see text).

421

Respiratory Distress

pH pH P1;co-2 P,co-1 Findings at NecropsyBefore After Before After Outcome CommentTHAM THAM THAM THAM Immediate Cause State of Lungs(mm. Hg) (mm. Hg) of Death7 03 7 12 - - Death IVH Normal; surfactant

46 hr. present6 90 7 04 95 64 Death RDS HMD Repeated apnoeic attacks

6 hr. IPPR from 4 hr.7 03 7 13 30 30 Death IN'H Normal; surfactant

41 hr. present7 00 7 09 105 73 Survived7 17 7 21 54 64 Survived Repeated apnoeic attacks6 88 6 92 43 39 Death INVH Good response to 2nd

18 hr. dose THAM; repeatedSmall pneumothorax;HMD

HMD; nosurfactant

Pneumonia; HMDHMD; no

surfactant

RDS; aspir.maternal blood

RDS

Subduralhaemorrhage

IVH

RDS

Tensionpneumothorax

Pneumonia

HMNID

H,MD

H,MD

BronchopneumoniaHMD

Atelectasis; pulm.haem.

HMD

Pulm. haem.; HMD

apnoeic attacksRepeated apnoeic attacksIPPR from onset

Good response to 2nddose of THAMNi

Massive antepartumhaemorrhage

Hypothermia; IPPRfrom 4- hr.

Repeated apnoeic attacksPlacenta praevia APHRepeated apnoeic attacks

Severe intrapartumasphyxia

Intrapartum asphvxiaIntrapartum asphyxia

RDS = respiratory distress syndrome. HMD atelectasis and hyaline membrane. * Pxo~. rise caused by brief apnoeic episode after infection.

clinical improvement corresponded with a sudden rise inP5o., which was sustained, and the infant subsequentlyrecovered.

Case 44. This infant was a vertex delivery followinga spontaneous premature labour. At 1 minute he wasgasping with a heart rate of 180'min. He subsequentlydeteriorated and required intubation but breathedspontaneously at 5 minutes. At 2 hours he had markedcostal recession, grunting respiration, and a respiratoryrate of 80/min. Air entry was poor. Chest x-ray filmwas compatible with RDS. Fig. 4 shows the blood gasand pH values before and after THAM. There was amarked improvement during the time THAM was beinggiven. The costal recession became less: the gruntingrespirations eased; and the infant showed more activity.At this time the right radial artery was catheterized andthereafter simultaneous arterial samples were taken fromthe radial artery and the lower aorta (i.e. above and belowthe ductus). It can be seen that PaO, values from eachsite were identical. A single dose of 5 mEq sodium

bicarbonate was given at the age of 6- hours as the pHhad not risen as expected. This caused a rise in pH buta fall in PaO.,. The infant made an uneventful recovery.

Fatal cases of 'bad prognosis' group. The associ-ated illness, the immediate cause of death andrelevant post-mortem lung findings are shown inTable II. Fig. 2 shows the PaO2 values (derivedfrom isopleths) at a P,o2 of 660 mm. Hg before andafter THAM.Three of these infants (Nos. 34, 36, and 39) had

substantial and sustained increases in the PaO2following THAM. They were recovering fromRDS and had normal gas tensions, when theysuddenly deteriorated and died at the age of 46, 41,and 18 hours, respectively. It will be noted thatNos. 34 and 36 had stable lungs at necropsy.

Infants Nos. 35, 38, 39, 40, 49, and 51 hadrepeated apnoeic attacks requiring intubation and

Effects of THAM in Respiratory Distress Syndrome

IVH

RDS

RDS

RDS

7 04

6 83

7 21

6 94

7 016 99

7*04

6 82

7 296 93

7 20

7 06

6 897 06

7 147 12

7 18

7 267 33

7 29

7 02

7 25

6 95

7 127 007 137 12

6 97

7 04

7 19

7 -09

7 -067 10

7 327 17

7 26

42

100

29

6560

16012759

120

30653838

8750626952

25

5224

32

6635

51

391277241

68

38

31

26

5862t373838

45

Death19 hr.

Death74 hr.

Death46 hr.

Death16 hr.

SurvivedSurvivedDeath

10 hr.Death

9 hr.SurvivedDeath

14 hr.Death

66 hr.Death

23 hr.SurvivedSurvivedSurvivedDeath

41 hr.Death

69 hr.SurvivedSurvived

Gupta, Dahlenburg, and Davisboo-500 -

400cr

EE 3000-o0-o6I 200-ECLE

~0 C-u

on-* 100 -I00

0-

Case 44

o . Il , __, ._ _._

2 3 4 5 8 14 ao 2632 38 44Aqe in hours

FIG. 4.-Case 44. Arterial oxygen tension, pH, and Pco2 before and after THAM (see text).

IPPR at times during their illness. Only one(No. 38) of this group survived: in the remainderthere was little response to THAM and, where a risein the P%o2 was obtained, it was not sustained.

It can be seen that of these 15 infants only 7 diedfrom RDS alone. Among these 7 infants, Cases 42and 55 showed satisfactory Pao2 values initially butsubsequently deteriorated, while infant No. 41 wastreated late in the disease, requiring IPPR from theoutset. The P%o2 did not rise with THAM in thesecases.

The pH which was below 7 * 20 at the outset in allcases rose following THAM in all but one; this rise,however, was not sustained.The PaCO2 was raised initially in all cases except

one. Again there was a consistent fall followingTHAM, but subsequently the levels once again rose.Infant No. 41 had the right radial artery catheterizedand simultaneous measurements of P5o2 were madefrom this site and from the lower aorta. The valuesobtained were identical in each instance, indicatingthere was no net right-to-left shunt through theductus. Cases 36 and 43 illustrate these points.

Case 36. This infant was born following a sponta-neous premature labour by vertex delivery. He wasapnoeic at 1 minute of age, heart rate 40/min., cyanosed,and limp. He was intubated and given IPPR. Thecolour improved and spontaneous respiration started at4 minutes, but he remained unresponsive for 15 minutes.He developed respiratory distress and at the age of 1 hourthe P.02 was 51 mm. Hg, and thepH 7 * 03 (Fig. 5). Tenminutes after THAM the Pao2 rose to 190 mm. Hg andthe pH to 7 - 13; the P.Co2 remained at 30 mm. Hg. Hisclinical condition also improved in that grunting

became less and costal recession eased. The gas tensionand pH remained satisfactory until the age of 37 hourswhen he suddenly became pale and limp. Shortlyafterwards he had the first of three apnoeic attacks anddespite IPPR he died at the age 41 hours.Necropsy showed a massive intraventricular haemor-

rhage and well-expanded lungs. Pressure-volumecurves were normal and surfactant was detected in thelungs.

Case 43. This infant was born following a sponta-neous premature labour at 33 weeks by a vertex delivery.At 1 minute the infant was gasping, heart rate 1 10/min.,and was blue. The colour improved with face-maskoxygen, but he then developed respiratory distress.Fig. 6 shows the blood gas tension and pH. At 3 hoursof age the Pao2 was 27 mm. Hg, pH 6-94, and P.co265 mm. Hg. Following THAM the Pao2 rose slightly,the PaCo2 fell, and the pH remained the same. Subse-quently doses of THAM produced little response andterminally the Paco2 rose and pH fell. (P5o0 measure-ments were not obtained in the latter stages due to afault in the apparatus.)

Necropsy showed the lungs to be collapsed; histologyrevealed hyaline membrane; and no surfactant wasdetected after 24 hours cycling. Owing to a technicalbreakdown this infant was not treated adequately in theearly stages of the disease.

Effect ofTHAM on arterial blood pressure.The arterial blood pressure was recorded directlyby a pressure transducer in 5 infants while THAMwas being administered via the umbilical vein.There was no change in the blood pressure duringthis time and for periods of up to 20 minutes after-wards.

-7*30

-7-20pH

- 7 -10

-700

422

Effects of THAM in Respiratory Distress Syndrome

7-4p0

-7-00E1 _= X X ___L__H

o2 6 12 18 24 30 36 42

Age in hours

FIG. 5.-Case 36. Arterial oxygen tension, pH, and Pco2 before and after giving THAM (see text).

Side effects. THAM has been shown to causehypoglycaemia and hyperkalaemia (Bennett andTarail, 1961) in animal experiments but no sucheffects were seen in this series of human infants.Respiratory depression (Nahas, Fink, Ploski, andTeneick, 1963) was observed on occasions, but onlyin the terminal stages of the illness and never withthe initial dose. On one occasion a fall rather thana rise in Pao2 was seen after an injection of THAMwhich was given into the umbilical artery. Thiswas not associated with a fall in PaCo2 or rise in pH.It will be noted that this patient had a tensionpneumothorax. We do not recommend the use ofTHAM in cases of predominantly metabolicacidaemia when the PaCO2 level is within the normalrange for newborn infants, since in these circum-stances it may cause respiratory depression.

Management. The use of THAM in respira-tory distress necessitates the presence of indwellingumbilical arterial and venous catheters over thecourse of the illness both for injection (which causestissue necrosis when given through smaller vessels)and for monitoring blood gas tensions and pH. Ameasure of the dramatic effect of treatment in somecases is the need to watch for excessively highoxygen tensions after THAM in treated infantsnursed in high ambient oxygen concentrations.The presence of indwelling catheters makes itimperative to avoid umbilical sepsis, and this can beachieved by spraying the cord stump at birth anddaily thereafter with 'polybactrin' and by paintingit with iodine before manipulation. The use ofheparin to keep the catheters open involves partialheparinization of the infant, and this may need to be

300aW

EE 200.

O-o.0~

0

E 100--o_

U

oli0

n

Case 43

pH

*690i 'I . 4-~~~~~~~

.I I2 3 4 5' 6 ; 8 9 10Age in hours

FIG. 6.-Case 43. Arterial oxygen tension, pH, and Pco2 before and after giving THAM (see text).

423

Gupta, Dahlenburg, and Daviscountered with appropriate quantities of protaminewhen the infant is tiny and frequent sampling isneeded.

In none of these cases was there any clinical orpost-mortem evidence of arterial thrombosis. In 3(very small) babies one leg became pale and pulse-less, but the circulation was restored immediatelyon removal of the catheter.

Finally, careful precautions have to be takenagainst bleeding from the cord stump when thecatheters are removed, and, indeed, from thecatheters themselves. In cases in which veryfrequent sampling is in the infant's interest, we donot hesitate to replace the blood that is removed by asmall transfusion, if this exceeds 10% of thecalculated blood volume.

DiscussionThe results presented re-emphasize the need for

a clear distinction between good prognosis and badprognosis RDS in the evaluation of treatment, theformer having a good prognosis as regards respira-tory sufficiency even without specific treatment foracidaemia, the latter doing badly if such therapy iswithheld. Initial PaO2 levels in 100% ambientoxygen concentrations have been shown by anumber of workers (Boston et al., 1966; Moss et al.,1965) to be a more reliable guide to prognosis thanchanges in pH, standard bicarbonate, or PaCO2levels; and on this basis it is possible to single out agroup of badly affected infants whose mortalityconventionally treated is of the order of 80%.Improvement in the outlook of this category is inour opinion likely to be a better guide to theefficiency of treatment than comparison of deathrates in consecutive series, or even in alternate butunselected cases, in a disease where the prognosis isaffected by so many variables. For instance, in ourexperience two-thirds of infants who die of massiveintraventricular haemorrhage (IVH) have no clinicalor pathological evidence of RDS, and in one-thirdof all cases dying of RDS there is an associated IVH,though in some of these the haemorrhage occurredwhen the infant no longer showed signs of respira-tory failure. The undoubted association of RDSand IVH may be coincidental on the basis of lowgestational age.

Treatment of the secondary metabolic effects ofrespiratory failure in RDS with mixtures of water,sodium, glucose, and bicarbonate, has been in voguesince its introduction by Usher (1959). From somecentres (Hutchison et al., 1964) it is claimed that theprognosis can be greatly improved thereby, thoughUsher himself has pointed out (1963) that glucose/

bicarbonate therapy does not seem to improve theprognosis in babies of under 1500 g. birthweight, acategory that includes many of the severe cases; andthe results of a reasonably well-controlled trial in alarge series of cases in Singapore have demonstratedno significant advantages to the treated infants(Teck, 1965).

In our view, death in RDS is primarily due tohypoxia and therefore, in a disease that lasts forupwards of 48 hours, any attempt to neutralize theacidaemia resulting from inadequate pulmonary CO2elimination by infusion of buffer or alkali is unlikelyto succeed in just those cases where respiratoryfailure is of a degree to threaten life. The demon-stration that in certain circumstances the administra-tion of alkali actually appears to improve oxygena-tion, as well as correcting pH, etc., and that thisimprovement may be sustained, is therefore ofconsiderable importance both in theory and practice.That it is seen alike in ultimately fatal cases, in babieson IPPR, and in bad risk infants who recover, makesit unlikely that the effect is due to spontaneousimprovement, to changes in respiration, or torandom fluctuation in Pao2. As suggested byGupta (1965) and independently by Chu et al. (1965),it also implies that part of the respiratory failure inRDS is the result of the inadequate pulmonaryperfusion; indeed, the very rapid response to treat-ment suggests that this may be the primary cause ofthe disease and that other changes, such as loss ofsurfactant, demonstrable at necropsy, may be insome measure secondary. Inadequate pulmonaryperfusion is presumably the result of either a lowcardiac output or of the shunting of blood pastfunctional lung tissue, i.e. through the foramenovale, through the ductus arteriosus, or through non-ventilated lung. This involves an alteration innormal pressure relationship (Rudolph, Drorbaugh,Auld, Rudolph, Nadas, Smith, and Hubbell, 1961;Moss, Emmanouilides, and Duffie, 1963) such that (1)in the case of the foramen ovale, right atrial pressuremust exceed left atrial pressure, (2) in the case of theductus, pulmonary arterial must exceed systemicarterial pressure, and (3) where there is perfusion ofunventilated lung, the vascular resistance in theunventilated segments must not increase. We havenot measured cardiac output, nor pressure in theatria, ventricles, or great vessels, but in 5 cases inwhich the arterial pressure was measured before andafter THAM, no change was noted. In 4 furthercases in which Pao2 levels were measured simulta-neously in the aorta (via the umbilical artery) and inthe right radial artery, there was no differencebetween Pao2 levels above and below the opening ofthe ductus arteriosus into the aorta. It is, therefore,

424

Effects of THAM in Respiratory Distress Syndrome 425unlikely that changes in cardiac output or in ductalshunting were responsible for the Pao2 changesobserved following THAM in these cases. Sincethe closure of the foramen ovale is dependent on achange in atrial pressure relationships caused by theincrease in pulmonary perfusion that takes place atbirth, and since this increase in pulmonary perfu-sion probably results from a fall in pulmonaryvascular resistance, it seems likely that in RDS theremay be a major shunting through the foramen ovaleand that this is in part due to a rise in pulmonaryvascular resistance producing a fall in pulmonaryblood flow. Pulmonary vascular resistance isknown to be dependent on pulmonary gas tensionsand pulmonary distension (Cassin et al., 1964;Rokseth, 1966), both of which are likely to bealtered in an adverse way in the course of RDS, andit, therefore, seems possible that the timely admini-stration of alkali may, by lowering pulmonaryvascular resistance, promote adequate pulmonaryperfusion, reduce shunting, improve blood andalveolar gas tensions, and prevent irreversiblepulmonary damage. It appears likely that thechanges in Pao2 described are due to a direct effectofTHAM in the lung or on the heart rather than tochanges in total body acid-base balance, the dosesused probably being too small to achieve the resultsobserved except through a concomitant improve-ment in respiratory function. It may be that insome cases lack of response to alkali was due to afailure of the drug to reach the lung or heart, eitherbecause of extreme shunting, lodgement of thevenous catheter in the liver, or injection into theumbilical artery instead of vein, whereas in othersit was because the drug was given too late whenirreversible pulmonary damage had occurred. Thiseffect could be on the pulmonary arteries (pulmon-ary artery pressure is known to be high except in theterminal stages of the illness (Rudolph et al., 1961;Moss et al., 1965)), but since the apparent pulmonaryhypoperfusion is accompanied by congestion of thelung at necropsy, an effect on the pulmonary veinsseems more likely. A. J. Hauck (1963, personalcommunication) has in fact observed a raisedpulmonary venous pressure in RDS in the presenceof a normal left atrial pressure. That hypoperfu-sion is the initiating disturbance would, for instance,explain the normal quantities of surfactant found atnecropsy in the two cases described in this report,which recovered from severe RDS and subsequentlydied with IVH after a period when there appearedto be no clinical signs of respiratory embarrassment.The evidence herein adduced, and the argumentbased thereon, suggests that RDS in its early stagesmay be aborted rather than ameliorated by the

administration of THAM, and that this effect ismore likely to be a primary one on the walls of thepulmonary vessels than secondary to changes intotal body acid base balance. Direct measurementsof the quantities involved (oxygen consumption,pulmonary arterial and atrial pressures), and thecalculation of pulmonary blood flow before and aftertreatment, would go far to disprove or substantiatethis hypothesis.

SummaryBabies with respiratory distress syndrome can

be divided into two groups with a good and badprognosis on the basis of an initial arterial oxygentension measurement taken when breathing 100%oxygen, providing a measure of the extent of V-Ashunting. Following intravenous THAM givenearly in the course of the illness, there is a consider-able and often sustained rise in arterial oxygentension in the bad prognosis group which can best beaccounted for by a diminution in such shunting.The alteration in calculated shunt induced byTHAM injection suggests that reversible hypoper-fusion is a factor in the aetiology of the pulmonarychanges seen in the respiratory distress syndrome.

The authors wish to acknowledge the help of Mrs.Lee, departmental secretary, Mr. J. F. Stevens, seniortechnician, colleagues at the Royal Postgraduate MedicalSchool, and Professor J. P. M. Tizard in the preparationof this paper. The work reported was stimulated byobservations on the effect of THAM in respiratorydistress made in the department of Paediatrics, Univer-sity of Groningen (Professor J. Jonxis, Dr. H. Visser, andDr. J. Troelstra). Dr. Colin Normand of UniversityCollege Hospital kindly carried out the determinations ofsurface tension-reducing activity and Dr. J. S.Wigglesworth the pressure volume curves. The labora-tory work was carried out in the Sir William CoxenLaboratories of the Nuffield Neonatal Research Unitusing instruments paid for by the Variety Club of GreatBritain.

REFERENcEsAstrup, P. (1956). A simple electrometric technique for the

determination of carbon dioxide tension in blood and plasma,total content of carbon dioxide in plasma and bicarbonatecontent in 'separated' plasma at a fixed carbon dioxide tension(40 mm. Hg). Scand. J. clin. Lab. Invest., 8, 33.

Bennett, T. E., and Tarail, R. (1961). The hypoglycemic activityof 2-amino-2 hydroxymethyl-1, 3-propanediol. Ann' N.Y.Acad. Sci., 92, 651.

Boston, R. W., Geller, F., and Smith, C. A. (1966). Arterial bloodgas tensions and acid-base balance in the management of therespiratory distress syndrome. J. Pediat., 68, 74.

Bradley, A. F., Stupfel, M., and Severinghaus, J. W. (1956). Effectof temperature on Pco2 and Po2 of blood in vitro. J7. appl.Physiol., 9, 201.

Cassin, S., Dawes, G. S., Mott, J. C., Ross, B. B., and Strang, L. B.(1964). The vascular resistance of the foetal and newlyventilated lung of the lamb. J. Physiol. (Lond.), 171, 61.

Chu, J., Clements, J. A., Cotton, E., Klaus, M. H., Sweet, A. Y.,Thomas, M. A., and Tooley, W. H. (1965). The pulmonaryhypoperfusion syndrome. Pediatrics, 35, 733.

426 Gupta, Dahlenburg, and DavisGupta, J. M. (1965). The effect of THAM on the oxygen tension

of arterial blood in neonatal respiratory-distress syndrome.Lancet, 1, 734.

-(1966). Studies in blood gas tension and acid base balance innormal and sick newborn infants. M.D. Thesis, University ofSingapore.

Hedley-Whyte, J., and Laver, M. B. (1964). 02 solubility andtemperature correction factors for Po2. J. appl. Physiol., 19,901.

Holmdahl, M. H., Nahas, G. G., Hassam, D., and Verosky, M.(1961). Acid-base changes in the cerebrospinal fluid followingrapid changes in the bicarbonate/carbonic acid ratio in theblood. Ann. N.Y. Acad. Sci., 92, 520.

Hutchison, J. H., Kerr, M. M., Douglas, T. A., Inall, J. A., andCrosbie, J. C. (1964). A therapeutic approach in 100 cases ofthe respiratory distress syndrome of the newborn infant.Pediatrics, 33, 956.

Moss, A. J., Emmanouilides, G., and Duffie, E. R., Jr. (1963).Closure of the ductus arteriosus in the newborn (Abstract).J. Pediat., 63, 709.

-, -, Retorri, O., Higashino, S. M., and Adams, F. H. (1965).Postnatal circulatory and metabolic adjustments in normal anddistressed premature infants. Biol. Neonat. (Basel), 8, 177.

Nahas, G. G. (1959). Use of an organic carbon dioxide bufferin vivo. Science, 129, 782.

-, Fink, B. R., Ploski, W. S., and Teneick, R. E. (1963). Thedepressant effects of tris (hydroxymethyl)-aminomethane andof mannitol on respiration. Ann. N.Y. Acad. Sci., 109, 783.

Nelson, N. M., and Reynolds, E. 0. R. (1964). Hyperbaric oxygenin patients with venoarterial shunts. Theoretical implications.New Engl. J. Med., 271, 497.

Peters, J. P. (1923). Studies of the carbon dioxide absorption curveof human blood. III. A further discussion of the form of theabsorption curve plotted logarithmically, with a convenient typeof interpolation chart. J. biol. Chem., 56, 745.

Rokseth, R. (1966). The effect of altered blood CO2 tension and pHon the human pulmonary circulation. Scand. J7. clin. Lab.Invest., 18, Suppl. 90.

Rosenthal, T. B. (1948). The effect of temperature on the pH ofblood and plasma in vitro. J. biol. Chem., 173, 25.

Rudolph, A. M., Drorbaugh, J. E., Auld, P. A. M., Rudolph, A. J.,Nadas, A. S., Smith, C. A., and Hubbell, J. P. (1961). Studieson the circulation in the neonatal period. The circulation inthe respiratory distress syndrome. Pediatrics, 27, 551.

Scopes, J. W., and Ahmed, I. (1966). Range of critical temperaturesin sick and premature newborn babies. Arch. Dis. Childh., 41,417.

Stevens, J. F., and Lanning, K. (1964). Micro Pco-2 equilibrationunit. Lancet, 2, 447.

Teck, T. W. T. (1965). Respiratory distress syndrome of thenewborn in Kandang Kerbau Hospital. J. Singapore paediat.Soc., 7, 44.

Tizard, J. P. M. (1964). Indications for oxygen therapy in thenewborn. Pediatrics, 34, 771.

Usher, R. (1959). The respiratory distress syndrome of prematurity.I. Changes in potassium in the serum and the electrocardio-gram and effects of therapy. ibid., 24, 562.- (1963). Reduction of mortality from respiratory distress

syndrome of prematurity with early administration of intra-venous glucose and sodium bicarbonate. ibid., 32, 966.

AppendixSources of Errors Due to Assumptions Made

in Calculating ShuntEnd pulmonary capillary oxygen content (Cco2) has

been derived from alveolar oxygen tension (PAO2) usingthe alveolar air equation. While entirely valid forsubjects who are breathing normally, the assumptiondoes not hold when there is any degree of respiratorydepression, for this causes a fall in P,&o2, as a result ofwhich the calculated value for Cco2 will be higher thanthe actual one; so giving a falsely high shunt value.However, if the alveolar Po2 is high, the fall in Cco2 due

to hypoventilation will be very small, and greater accuracycan therefore be expected in calculating the right-to-leftshunt if the subject breathes a high concentration ofoxygen rather than air (Perkins, Adams, Flores, Harper,and Landahl, 1958).

Large right-to-left shunts (over 70%) cause arterialPco2 appreciably to exceed alveolar values (Strang andMacLeish, 1961), and this will give a falsely low value foralveolar Po2 by the alveolar air equation. As the erroris small no correction for it has been made.

For the calculation of alveolar Po2 the respiratoryquotient has been assumed to be 0 *8. In the newbominfant RQ varies from 0 7 to 1-0 (Cross, Tizard, andTrythall, 1957) and this could result in a maximum errorof 40 mm. Hg in the calculated PAO2 in a severely illinfant. Such infants, however, will be breathing veryhigh concentrations of oxygen and as such the error inthe calculated shunt will be small.As explained in the text, mixed venous samples are diffi-

cult to obtain in newborn babies. Strang and MacLeish(1961) observed that the arteriovenous oxygen differencesin all their cases were between 2 and 6 ml./100 ml., andthey showed that the calculated shunt value at 4 did notdiffer from those at 2 and 6 ml./100 ml. by more than 12%;hence, with an assumption of an arteriovenous differenceof 4 ml./ 100 ml. an error of 12% may occur.Oxygen capacity has been assumed to be 20 ml./100 ml.,

which is equivalent to 15-8 g. Hb/100 ml. This is, ofcourse, not correct for all infants since their Hb con-centrations may differ. Moreover, because of thechanges in haematocrit that occur in the first few days oflife (Gairdner, Marks, Roscoe, and Brettell, 1958) thevalue in a particular infant will also vary. To determinethe magnitude of the error, shunts were calculated byassuming oxygen capacities of 18 -2 and 23-4 ml./100 ml.(i.e. Hb concentrations of 14 and 18 g./100 ml.). Themaximum difference was 10% from those calculatedby assuming a value of 20 ml./100 ml.A further uncertainty is introduced by the fact that the

proportion of foetal Hb may vary in the infants studied;whereas the standard oxygen dissociation curve forfoetal Hb at 370 C. atpH 7 *40 was used to determine theoxygen content in all cases.Another possible source of error is that the results are

expressed as percentage of cardiac output, which ofcourse may vary from time to time in a particular infant.No correction has been applied to the oxygen dissocia-tion curve for differences in temperature and pH, asrecommended by Bradley, Stupfel, and Severinghaus(1956), since the error introduced by this omission isinsignificant in relation to other possible errors as aresult of the assumptions made.

Because of all these possible sources of error thechanges in shunting demonstrated must be considered asonly semi-quantitative, but this does not affect theargument of the paper, which remains valid.

REFERENCES

Bradley, A. F., Stupfel, M., and Severinghaus, J. W. (1956). Theeffect of temperature on Pco2 and Po2 of blood in vitro. J. appl.Physiol., 9, 201.

Effects of THAM in Respiratory Distress Syndrome 427Cross, K. W., Tizard, J. P. M., and Trythall, D. A. H. (1957). The

gaseous metabolism of the newborn infant. Acta paediat.(Uppsala), 46, 265.

Gairdner, D., Marks, J., Roscoe, J. D., and Brettell, R. 0. (1958).The fluid shift from the vascular compartment immediatelyafter birth. Arch. Dis. Childh., 33, 489.

Perkins, J. F., Jr., Adams, W. E., Flores, A., Harper, P. V., andLandahl, H. D. (1958). Arterial 02 saturation vs. alveolar 02tension in anatomical venous-arterial shunting. J. appl.Physiol., 12, 71.

Strang, L. B., and MacLeish, M. H. (1961). Ventilatory failure andright-to-left shunt in newborn infants with respiratory distress.Pediatrics, 28, 17.

The following articles will appear in future issues of this Journal:Adrenal Cortex in Marasmic Children. By Samir S. Najjar and John G. Bitar.Effects of Amino Acid Loads on a Healthy Infant with the Biochemical Features of Hartnup Disease. ByJ. W. T. Seakins and R. S. Ersser.Immunochemical Estimation of Some Proteins in Nigerian Paired Maternal Foetal Blood. By H. McFarlane andI. 0. K. Udeozo.Treatment of Classical Phenylketonuria. By M. S. McBean and J. B. P. Stephenson.Duodenal Ulcer in Childhood. A Study of Predisposing Factors. By B. F. Habbick, A. G. Melrose, andJ. C. Grant.Milk Intolerance in Infancy. By Hilton Silver and D. M. Douglas.Classification of Protein-calorie Undernutrition in Children. By K. L. Mukherjee.Comparison of Methods for Evaluating Surface Properties of Lung. By Gillian Gandy, J. G. Bradbrooke,B. T. Naidoo, and Douglas Gairdner.Transient Respiratory Distress Syndrome in the Newborn. By John J. Downes, Subhash Arya, Grant Morrow,III, and Thomas R. Boggs, Jr.Intracranial Haemorrhage Associated with Hyaline Membrane Disease. By V. C. Harrison, H. de V. Heese, andM. Klein.Incidence and Treatment of Infantile Gastro-enteritis in General Practice. By David Wheatley.Myoclonic Epilepsy in Childhood. By J. R. Harper.Angiokeratoma Corporis Diffusum. Some Clinical Aspects. By A. W. Johnston, S. D. V. Weller, andB. J. Warland.Emetine Hydrochloride and Dehydroemetine Combined with Chloroquine in the Treatment of Children withAmoebic Liver Abscess. By J. N. Scragg and S. J. Powell.Serum-Insulin Changes Following Administration of L-leucine to Children. By D. B. Grant.Lincomycin in the Treatment of Penicillin-resistant Staphylococcal Infections in Children. By J. F. R. Bentleyand D. Pollock.Hereditary Pituitary Dwarfism with Spontaneous Puberty. By M. Seip, C. B. van der Hagen, and 0. Trygstad.Histidinaemia: A Child and his Family. By A. R. R. Cain and J. B. Holton.

7