Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1087

_____________________________ _____________________________

Urea and Non – Protein NitrogenMetabolism in InfantsWith Special Reference to the Sudden

Infant Death Syndrome (SIDS)

BY

MARY GEORGE

ACTA UNIVERSITATIS UPSALIENSISUPPSALA 2001

Dissertation for the Degree of Doctor of Medical Science in Anaesthesiology and IntensiveCare presented at Uppsala University in 2001

ABSTRACT

George, M. 2001. Urea and Non – Protein Nitrogen Metabolism in Infants. With SpecialReference to the Sudden Infant Death Syndrome (SIDS). Acta Universitatis Upsaliensis.Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1087, 43pp. Uppsala ISBN 91-554-5141-1.

A large amount of non – protein nitrogen, in the form of urea and ammonium, is present inhuman breastmilk; however its physiological role in the infant is as yet not fully understood.It has been hypothesized that an insufficient enteric metabolism of urea could play a role inthe sudden infant death syndrome (SIDS). This thesis was undertaken to study the entericmetabolism of non – protein nitrogen in healthy infants, in comparison with those who hadsuccumbed to SIDS.

Aerobic and anaerobic faecal microflora, were studied in healthy infants from birth to 6months of age. They were found to appear in faeces within 3 days of birth and were presentthroughout the first 6 months of life. The effect of nitrate, nitrite and nitric oxide on faecalurease activity was investigated and found to be inhibitory in action. The sigmoid faecalurease activity and sigmoid faecal urea content of SIDS infants were compared to those ofcontrol infants; significantly lower sigmoid faecal urease activity and greater sigmoid faecalurea content were found in the SIDS infants. The total number of SIDS cases occurring inSweden during the period 1990 through 1996 was analysed regarding geographical andseasonal distribution, in relation to the nitrate concentration in drinking water and changes inthe groundwater level. The northernmost parts of the country had its highest incidence whenthe rest of the country had its lowest incidence, and the occurrence of individual deaths wasassociated with the recharge of groundwater, which is known to increase its nitrate content.The effect of ingested ammonium on carbon dioxide production was determined in healthyinfants using the doubly labelled water technique. No change in carbon dioxide productionwas observed.

An insufficient enteric metabolism of urea in infants and peak or greatly varying nitrateconcentrations in drinking water are associated with the occurrence of SIDS. Ingestedammonium supplements in the given doses did not influence carbon dioxide production inhealthy infants.

Key words: Urea, non – protein nitrogen metabolism, SIDS, drinking water nitrateconcentration, carbon dioxide production.

Mary George, Department of Surgical Sciences – Unit for Anaesthesiology and IntensiveCare, University Hospital, SE – 751 85 Uppsala, Sweden

© Mary George 2001

ISSN 0282-7476ISBN 91-554-5141-1

Printed in Sweden by Eklundshofs Grafiska, Uppsala 2001

There is no finer investment for any community than putting milk into babies.

Winston Churchill 1943

This thesis is based on the following papers, which are referred to in the text by their Romannumerals:

I Faecal microflora and urease activity during the first six months of infancy. George M, Nord CE, Ronquist G, Hedenstierna G, Wiklund L. Upsala J Med Sci 1996, 101: 233 - 250.

II Sudden infant death syndrome and nitrogen metabolism: further development of a hypothesis. Wiklund L, George M, Nord CE, Ronquist G, Saldeen T. Eur J Clin Invest 1998, 28: 958 - 965.

III Incidence and geographical distribution of sudden infant death syndrome (SIDS) in relation to content of nitrate in drinking water and groundwater levels. George M, Wiklund L, Aastrup M, Pousette J, Thunholm B, Saldeen T, Wernroth L, Zarén B, Holmberg L. Eur J Clin Invest Accepted for publication.

IV The effect of ammonium chloride ingestion on carbon dioxide production in infants determined by using the doubly labelled water technique. George M, Forsum E, Wright A, Coward WA, Wiklund L. Submitted for publication.

CONTENTS

INTRODUCTION 7

Urea metabolism and acid - base balance 7Sudden infant death syndrome (SIDS) 8Hypothesis 10

AIMS OF THE THESIS 12

SUBJECTS AND METHODS 13

Subjects 13Paper I and II Analyses of faecal specimens 13

Bacteriological analyses 13 Biochemical analyses 14

Paper III Geographical and seasonal distribution of SIDS 17Paper IV Doubly labelled water technique (DLW) 18

RESULTS 22

Paper I 22Paper II 24Paper III 25Paper IV 27

DISCUSSION 29

Urea metabolism 29Breastfeeding 30Postmortem vitreous humor levels of urea nitrogen 31Preceding illnesses 31Helicobacter pylori infection 31Carbon dioxide production and total energy expenditure 32Smoking 32Prone position and body temperature 33Increased winter incidence of SIDS 34

CONCLUSION 36

Future research 36

ACKNOWLEDGEMENTS 38

REFERENCES 39

ABBREVIATIONS

2H Stable isotope, deuterium

18O Stable isotope

CO2 Carbon dioxide DLW Doubly labelled water

H+ Hydrogen ion

HCO3- Bicarbonate ion

NaCl Sodium chloride

NAD+ Nicotinamide adenine dinucleotide

NADH Dihydronicotinamide adenine dinucleotide

NH4+ Ammonium ion

NO Nitric oxide

SIDS Sudden infant death syndrome

7

INTRODUCTION

As early as 1949 it was known that human breastmilk contains large amounts of non - protein

nitrogen in the form of urea and ammonium. (1) Non - protein nitrogen, half of which is made

up of urea nitrogen, accounts for 25 per cent of the total nitrogen in breastmilk. (2,3,4) The

physiological role of this urea and ammonium and its contribution to normal metabolism in

the infant is as yet not fully understood. In the adult, urea is considered to be a water soluble

relatively nontoxic waste product, which is excreted as urine through the kidneys.

Ingested urea, contained in breastmilk, is metabolized by urease - containing bacteria, which

are found in the intestines within 5 days of birth. (5,6,7,8) Currently available infant formula

feeds, which are similar to breastmilk in their high lactose content and low protein content,

were found to stimulate growth of intestinal Bifidobacteria but unable to suppress pathogenic

bacteria such as Clostridia and Klebsiella. (9) In adults, urea produced in the liver enters the

enterohepatic circulation and diffuses into the intestine, where about 20 per cent of the urea is

hydrolyzed by intestinal microflora into ammonia while the rest of it is excreted unchanged in

the urine. (10) The therapeutic use of antibiotics such as neomycin to abolish intestinal

bacterial action on urea in patients with impending hepatic coma is well established.

In the infant, the ammonia from dietary urea is transported to the liver where it is metabolized

once again to urea and excreted through the kidneys. Some of the nitrogen from dietary urea

is incorporated into body proteins. (11, 12, 13)

8

Urea Metabolism and Acid - Base Balance

Traditional view

Ingestion of protein results in the release of amino acids into the splanchnic circulation. The

deamination of amino acids, which is the removal of the amino groups from the amino acids

occurs mainly in the liver. The ammonium ion together with bicarbonate is synthesized into

urea via the urea cycle. Urea is then excreted through the kidneys in the urine. (14)

In the proximal tubular cells of the kidneys, deamination of glutamine yields 2 ammonium

ions and 1 α-ketoglutarate ion; the ketoglutarate ion is metabolized to glucose, or to carbon

dioxide and water producing 2 bicarbonate ions. Thus a bicarbonate ion is added to the blood

for each hydrogen ion that is excreted in the form of an ammonium ion. In the Loops of Henle

and in the medullary interstitium the concentration of the ammonium ion increases by means

of a countercurrent multiplication before being secreted into the collecting ducts. (15)

Alternative view according to Atkinson and Camien

Catabolism of protein produces a large amount of bicarbonate ion (base) rather than acid as is

generally believed. Bicarbonate ions are produced from the α carboxyl group of an amino

acid and from the carboxyl or carboxamide group of glutamate, aspartate, glutamine and

aspargine as a result of amino acid metabolism. Approximately 1 mol of bicarbonate ion is

produced from a daily consumption of 100 g of protein. Bicarbonate ions combine with

ammonium ions synthesizing urea. Synthesis of urea supplies protons, which titrate the large

amount of bicarbonate produced, thus maintaining acid - base balance.(16)

2NH4+ + 2HCO3

− NH2 − CO − NH2 + CO2 + 3H2O

9

Sudden Infant Death Syndrome (SIDS)

SIDS is defined as the sudden death of an infant or young child, which is unexpected by

history, and in whom a thorough post - mortem examination fails to demonstrate an adequate

cause of death.

Sudden and unexpected deaths in infancy have been described to occur as early as 500 B. C.

(17) Up until the early nineteenth century, overlaying of infants by adults was considered to

be the cause of such infant deaths. The mother or wetnurse sharing the infant`s bed was

forbidden; if death occurred, it was punishable by law in the middle ages, both in Sweden and

in other European countries.

In the nineteenth century, several theories were proposed and subsequently discarded. Such

was the fate of the theories known as the status thymico- lymphaticus (large thymus),

hypogammaglobulinemia, overwhelming viremia, hypoparathyroidism, unstable myocardial

conduction systems and cow`s milk hypersensitivity. (18)

A significant step forward was the proposition of the apnoea hypothesis in the 1970`s, (19,

20) and this led to the widespread use of apnoea monitors. Naeye, a pathologist, found certain

characteristic tissue markers for hypoxia and hypoxemia in SIDS cases, thus supporting the

apnoea hypothesis. (21) Histological studies of the central nervous system were conducted in

the 1980`s, and a defect in the brain stem was suggested as the cause of the respiratory failure

of SIDS. However, no etiological factor was found. (22)

The findings of a major epidemiological study of SIDS was published in 1988. (23) The

significant maternal factors were found to be cigarette smoking during pregnancy, single civil

status, age less than 20 years at first pregnancy, anaemia, use of illegal drugs and low weight

gain during pregnancy; the factors in the infants were found to be low birth weight, frequent

illnesses and prolonged post - natal stay in the nursery.

10

Adoption of the supine sleeping position instead of the prone position, as well as propaganda

encouraging breastfeeding in the 1990`s, led to a pronounced drop in the incidence of SIDS in

several countries worldwide. (24, 25, 26, 27)

The etiological factor or factors of SIDS have yet to be elucidated.

Hypothesis Proposed by Wiklund

Large amounts of non - protein nitrogen in the form of urea and ammonium ions are found in

human breastmilk. (1, 4) Bifidobacteria and other anaerobic bacteria, which are found in the

gut of both breastfed and formulafed infants (9, 28) produce urease (6,7) which metabolizes

ingested urea to ammonium and bicarbonate ions. The ammonium ions are absorbed and

incorporated into protein in the infant (11, 12, 29, 30, 31, 32) or metabolized into urea. (30)

The synthesis of urea is accompanied by the production of hydrogen ions which combine with

endogenously produced bicarbonate ions resulting in the formation of carbon dioxide and

water. (33)

Normal Infants

Gut Bifidobacteria (gut) Urea (ingested) NH4

+

urease producing

Liver NH4

+ (from gut) Urea + H+

Blood H+ + HCO3

− CO2 + H2O (from other (maintains metabolism ) acid- base balance )

Figure 1. Metabolism in normal infants

11

According to the hypothesis proposed by Wiklund in 1996 (34) a derangement of the above

metabolism could lead to progressive alkalosis and eventually death as seen in SIDS infants.

SIDS Infants

Deficient bacterial breakdown of urea Decreased production of NH4− ions

(gut)

Decreased production of H+ ions (liver)

Accummulation of HCO3− ions

Metabolic alkalosis Compensatory hypoventilation Accummulation of CO2 Carbon dioxide narcosis Hypoxia and death

Figure 2. Hypothetical mechanism in SIDS

To verify this hypothesis the following research was conducted, and the studies undertaken

form the basis of this thesis.

12

AIMS OF THE THESIS

To elucidate the role of the urea and nitrogen metabolism in healthy infants and a possible

causal role of a deficient metabolism in SIDS infants.

The specific aims of the studies were:

1. To study the faecal microflora and urease activity in healthy infants during the first six

months of life.

2. To compare the faecal microflora, urea content and urease activity in SIDS infants with

those of infants who died unexpectedly of other causes.

3. To determine the incidence and geographical distribution of SIDS in relation to the nitrate

content of drinking water and groundwater levels.

4. To study the effect of ingested ammonium chloride on carbon dioxide production and

total energy expenditure in healthy infants using the doubly labelled water technique.

13

SUBJECTS AND METHODS

Subjects

Paper I. Thirteen fullterm healthy infants were followed up from birth until 6 months of age.

Faecal specimens, which were collected during the first 3 days following birth, at 2 months, at

3 months and at 6 months, were examined bacteriologically, and biochemically for urease

activity and urea content.

Paper II. Thirty consecutive cases of unexpected infant death autopsied forensically were

studied. Twenty-two of these infants were found to have succumbed to SIDS and the

remaining 8 to other causes. Their faecal specimens were examined for bacteriological

microflora, urease activity and urea content and compared.

Paper III. The total number of SIDS deaths (636 cases) occurring in Sweden during the

period 1990 - 1996 was analysed regarding their geographical and seasonal distribution, in

relation to the nitrate content of drinking water and changes in groundwater levels.

Paper IV. In 8 fullterm healthy infants, carbon dioxide production and total energy

expenditure were determined, using the doubly labelled water (DLW) technique, following

ingestion of ammonium supplements, and compared to when no additional ammonium was

ingested. Urine pH, urea and ammonium concentrations were also analysed and compared.

Method

Paper I and II

Analyses of Faecal Specimens

Faecal specimens were collected after spontaneous defecation (Paper I), and from the sigmoid

colon ( Paper II).

Bacteriological Analyses: Small quantities of faeces from each infant were cultured

aerobically and anaerobically and then analysed for the presence of microflora. In addition, in

Paper I, the urease activity of each bacterial species was determined after incubation of the

14

specific bacterial species for 2 h in the presence of urea (0.3mol/L) and a bacteriological

substrate at pH 6.8 - 6.9.

Biochemical Analyses: Urea and ammonium concentrations, urease activity and gaseous

concentration of nitric oxide (NO)

Four faecal specimens each weighing approximately 40 mg were harvested from each

individual and suspended in 4mL of sterile physiological NaCl solution in sterile test tubes.

Two of the suspensions were preincubated anaerobically and the other two aerobically for 36h

at 33oC, followed by centrifugation to separate faecal material from a clear supernatant.

Before centrifugation the gas in the aerobic tubes was transferred into a syringe for immediate

analysis of NO. The anaerobic preincubation was carried out in an anaerobic tight chamber

containing a gas mixture of 90% CO2 and 10% H2 circulated over a palladium catalyst

(GasPak Plus, Becton Dickinson Microbiology Systems, Cockeysville, MD, USA) while

aerobic preincubation was done in ambient air sealed test tubes. (Figure 3)

4 faecal specimens (40mg) 2 Aerobic Incubated at 33o Csuspended in 4mL NaCl 2 Anaerobic for 36 h (90% CO2 and 10% H2 over a palladium catalyst

Aerobic tube gas NO analysis

Centrifugation Supernatant + Faecal mixture

Figure 3. Preanalytical procedure

After testing the pH of the supernatants (2 from anaerobic conditions and 2 from aerobic

conditions), they were subdivided into 400 µL (600 µL in Paper II) samples (in duplicate),

each of which was mixed with 3mL of sterile physiological NaCl solution forming an

incubation mixture. Hence 8 incubation mixtures (4 from anaerobic conditions and another 4

15

from aerobic conditions) were supplemented with urea to a final concentration of 1.95±0.12

(1.91±0.4 in Paper II) mmol/L, and 8 additional incubation mixtures to a final urea

concentration of 3.40±0.22 (3.38±0.47 in Paper II) mmol/L for subsequent incubations. Four

other incubation mixtures (2 from anaerobic conditions and 2 from aerobic conditions) were

incubated without any additional urea in order to determine the endogenous faecal content of

urea. (Figure 4) Finally to 4 test tubes containing sterile, physiological NaCl solutions, urea

was added to obtain final concentrations as stated above (2 of each concentration) and they

were also incubated in the same way (blanks). Each individual generated a total of 24 tubes

which were incubated for 30h at 33oC before analysis.

Urease activity

400µL supernatant (+3mL NaCl) 2 aerobic samples +urea urea 1.91±0.4 mmol/L

2 aerobic samples +urea urea 3.38±0.47 mmol/L

2 anaerobic samples +urea urea 1.91±0.4mmol/L

2 anaerobic samples +urea urea 3.38±0.47mmol/L

Urea content

400µL supernatant + 3mL NaCl 2 aerobic samples

2 anaerobic samples

Figure 4. Incubation mixtures

Nitric oxide (NO) concentration

The NO concentration of the gaseous atmosphere of the aerobic tubes was determined in a

chemiluminescence NO − NO2 −NOX analyser (Model 42, Thermo Environmental Instuments

Inc, USA) and expressed in parts per billion (ppb).

All incubation mixtures were analysed regarding their contents of urea and ammonium ions in

a Hitachi 911 analytical system, each value being the mean of 4 analyses. The ammonium

16

ions thus formed were reacted in a reductive amination of α - ketoglutarate catalysed by

glutamate dehydrogenase with concomitant oxidation of NADH, which was followed

spectrophotometrically at 340 nm. Hence stoichometric relationships existed between the

amounts of NADH converted and ammonium ions formed (as a function of the actual urea

concentration) in the urease reaction. Ammonium ions in the samples were assayed similarly

but without urease. (Figure 5) The urea and ammonium ion concentration were expressed in

mmols and µmols per gram (wet weight) faeces respectively, after blank corrections. All

chemicals were of pro analysi grade from Boehringer - Mannheim, Germany.

urease Urea CO2 + 2 NH4

+

NADH + H+ NAD+

NH3 L Glutamate α Ketoglutarate Dehydrogenase L Glutamate

Figure 5. Urea and ammonium biochemical reactions

Statistics

Biochemical analyses were presented as means (SD) and range, while results of the

quantitative bacteriological cultures were given as box plots displaying the 10th, 25th, 50th

(median), 75th and 90th percentiles. In Paper II the ammonium ion production and urea content

in the faecal samples of the SIDS infants were compared with those of the control infants

using the Mann Whitney U test. Standard Pearson regression analysis was used to determine

correlations between the faecal urea content, the faecal ammonium ion production and the

faecal NO production. Values p<0.05 were considered significant.

17

Paper III

Geographical and Seasonal Distribution of SIDS

All of the 636 SIDS cases which occurred in Sweden from 1990 through 1996 was analysed

regarding their geographical and seasonal distribution in relation to nitrate concentrations in

drinking water and groundwater level changes. Copies of the death certificates were obtained

from Statistics Sweden (Statistika Centralbyrån). For the years 1990 - 1996 annual population

data from the entire country, for each of the counties and muncipalities and also monthly

statistics from each muncipality regarding births and deaths of infants in the age group 1 - 11

months were provided by Statistics Sweden.

Addresses of the infants recorded on the death certificates were translated to geographical

coordinates by the National Land Survey of Sweden; this agency also supplied information

regarding source of water supply, whether from a private well or from the muncipal water

works, to the recorded addresses.

The Geological Survey of Sweden (SGU) matched the geographical coordinates with

chemical analyses of groundwater data from its National Groundwater Network records. The

SGU record chemical analyses of groundwater twice a year (75 stations) and groundwater

levels twice a month (400 stations). These data were used in comparing timing of individual

infant deaths with groundwater levels in the county of residence of the infants. The nitrate

concentrations of water of private wells (25000 wells) are analysed approximately 1 - 3 times

per 10 year period and are used to produce maps of the nitrate concentrations of groundwater;

these maps were used to estimate nitrate concentrations of drinking water. The nitrate content

of water supplied by the muncipal water works are recorded regularly and this information

was available to us.

Statistics

18

A spatial scan statistic (SaTScan) developed by Kulldorff (35) was used to detect clusters of

rates of SIDS cases in relation to live births in each of Sweden´s municipalities. SIDS cases

were assumed to be Poisson distributed and the maximum cluster size assumed to be 50% of

the total population. Each geographic centre was considered to be the centre of a potential

cluster. Likelihood functions were evaluated using a circle of variable size and location to

scan the whole map for areas with high and low rates. Monte Carlo simulation methods were

used to approximate the distribution of the likelihood function, and p values for the most

likely cluster were determined after compensation for the changing incidence of the disease.

Poisson regression analysis was performed to assess if there was a seasonal variation of SIDS

cases in different parts of the country and this was analysed both according to the months of

the year and according to the quarters of the year.

Chi square tests were used to compare the proportion of days when the groundwater level was

high with that when it was low. Groundwater was considered high when it was increasing or

at least 75% of the annual maximum level, and low when it was decreasing or below 75% of

this level. The expected number of SIDS cases was derived by multiplying the total number of

SIDS cases by the proportion of days when the groundwater level was high.

Spearman rank correlation was used to determine if there was a relationship between the local

incidence of SIDS and the nitrate concentration in the public water supply.

Significance levels of 0.05 were used throughout the study.

Paper IV

Doubly Labelled Water Technique (DLW)

This technique is safe and non - invasive with an accuracy of ±5%. It is robust and is

especially useful in noncompliant subjects such as infants. This method not only measures

carbon dioxide but may be used to derive total energy expenditure, total body water and water

19

metabolism. However, the cost and the technical expertise required for analysis limit the use

of this technique.

Principle

An accurately weighed dose of DLW solution containing the stable isotopes deuterium 2H and

18O is given to the subject orally. 2H mixes with the water pool while 18O mixes with both

water and bicarbonate pools, which are in equilibrium through the carbonic anhydrase

reaction. (36)

CO2 + H2O H2CO3 HCO3

− + H+

carbonic anhydrase Figure 6. Carbonic anhydrase reaction

The rate constant for 2H disappearance measures water flux, while the rate constant for 18O

disappearance measures the sum of water and carbon dioxide (CO2) flux; thus the difference

between the rate constants gives the CO2 flux. (37)

KD ≅ H2O flux

KO ≅ H2O + CO2 flux

∴ KO − KD ≅ CO2 flux

Figure 7 . Rate constants for 2H (KD ) and 18O (KO )

2H and 18O enrichments were determined by mass spectrometric analysis (Micromass,

Wythenshawe, UK) of urine samples collected before dosing with DLW (predose) and

serially for up to 12 days after dosing. Rate constants were derived from the slopes of lines

plotted with isotope enrichment (y - axis) against time (x - axis). (37)

An accurately weighed dose of DLW containing 96 mg 2H2O and 240 mg H218O per kg body

weight was given via a nasogastric tube to 8 healthy fullterm infants. Urine samples were

20

collected prior to DLW dosing (predose) and then following the DLW ingestion, at 4 hours,

on day 2, 6, 10 and 12. The samples were collected, using self - adhesive urine bags, and then

stored at − 20oC until analysis using the mass spectrometer. (Figure 8)

IsotopeEnrichment 2H

18O

1 2 6 10 12 Days

Figure 8. Schematic diagram of isotope enrichment plotted against time ( days)

Carbon dioxide production and total energy production were compared between the 2 study

periods i. e. with and without ammonium supplementation.

Ammonium Supplements

Eight infants were randomized to receive either supplements or no supplements during the

first study period (Part A ) consisting of 12 days, and vice versa during a second study period

(Part B) also of 12 days duration; Part B was performed after a minimum of 2 weeks after

completion of Part A.

A sterile water solution containing 500 µmol/mL of ammonium chloride (Addex -

Ammoniumklorid Fresenius Kabi, Uppsala, Sweden) was prepared. The mothers were

instructed on how to measure accurately the assigned doses, to be given to their infants, in 2

mL syringes. An oral dose of 400 µmol/kg/day was given as divided doses of 100 µmol/kg 4

times daily through the 12 days of the study period.

21

Biochemical Analyses of Urine

Urine pH, urea and ammonium concentrations were determined as described previously in

Paper I and II. The values obtained following ingestion of ammonium supplements were

compared to those when no ammonium supplements were ingested.

Statistics

Mean and SE were calculated for age and weights of the infants, carbon dioxide productions,

TEE, 2H and 18O dilution spaces and elimination rate constants. Wilcoxon signed rank test

and Mann - Whitney U test were used for comparing carbon dioxide productions, TEE, urine

urea, ammonium concentrations and pH, with and without ammonium supplementation.

22

RESULTS

Paper I

Bacteriological Analyses

Anaerobic microflora

Bifidobacteria and Bacteriodes were the most prevalent species and were present throughout

the first six months of life.

Aerobic microflora

Streptococci, Enterococci, Escheria coli and Klebsiella were the commonest species and were

prevalent throughout the first six months.

Faecal pH

The pH of the anaerobic preincubation mixtures was approximately 0.2 pH units lower than

that of the aerobic samples, and this remained so throughout the 6 months. The mean (SD) of

the pH of the anaerobic and aerobic samples were 5.3 (0.6) and 5.6 (0.8) respectively at 3

days after birth and 5.7 (0.7) and 5.9 (0.6) respectively at six months.

Biochemical Analyses

Faecal urease activity

The faecal urease activity expressed as ammonium ion formation per gram wet weight of

faeces after 30 h incubation, decreased steadily from 3 days after birth until six months of age.

(Table 1)

23

Aerobic preincubation Anaerobic preincubation Incubationureaconcentration(mmol/L)

1.95± 0.12 3.40 ± 0.22 1.95 ± 0.12 3.40 ± 0.22

3 days 56 ± 68 67 ± 74 40 ± 34 60 ± 342 months 40 ± 60 57 ± 79 36 ± 46 47 ± 563 months 36 ± 67 42 ± 72 22 ± 26 32 ± 356 months 15 ± 14 20 ± 17 11 ± 8 12 ± 9

Table 1. Faecal urease activity expressed as ammonium ion formed (mmol/g faeces wetweight after 30 h incubation) ± SD in infants from 3 days to 6 months of age, using a low andhigh urea concentration, under aerobic and anaerobic preincubations.

Faecal urea content

The faecal urea content was found to be 5 - 15 µmol/g faeces wet weight from 2 to 6 months

of age, and approximately 40 µmol at 3 days after birth. (The water content of the faeces was

assumed to be 90%) (Table 2)

Age Urea content µmol/g faeces (wet weight)3 days 39 ± 572 months 4 ± 12 3 months 5 ± 176 months 6 ± 14

Table 2. Faecal urea content ± SD of infants from 3 days to 6 months of age.

Nitric oxide concentration (NO)

The NO concentration in the aerobically preincubated test tubes increased from 52 ± 35 ppb

in 3 day old infants to 575 ± 419 ppb (mean ± SD) in 6 month old infants. (Table 3)

NO, sodium nitrate and sodium nitrite, when added to preincubation mixtures were found to

have inhibitory effects on faecal urease activity. However this inhibition was not seen when

added to sterile urea solutions with jack bean urease. Thus, suggesting that the bacterial

production of urease is inhibited and not the urease activity itself.

24

Age Nitric oxide concentration (ppb)3 days 52 ± 352 months 151 ± 623 months 149 ± 866 months 575 ± 419

Table 3. Faecal concentrations of nitric oxide ± SD in aerobic preincubated test tubes ininfants 3 days to 6 months of age.

Paper II

Out of a total of 30 cases of unexpected infant deaths, 22 were considered to be due to SIDS

and 8 due to other causes (pneumonia, shaken baby syndrome, suffocation, septicaemia,

subendocarditis, myocarditis). The 8 infants were referred to as controls and were compared

with the SIDS infants in the study.

Bacteriological Analyses

In both the SIDS cases and in the controls the commonest aerobic microflora were

Enterococci and Escheria coli, while the commonest anaerobic microflora were

Bifidobacteria, Clostridia and Bacteriodes.

Biochemical Analyses

Faecal urease activity

The faecal urease activity expressed as ammonium ions formation (mmol/g wet weight of

faeces) after 30h at 33oC with urea as substrate at two different concentrations was found to

be significantly lower in the SIDS infants compared to the controls. (p<0.05) (Table 4)

25

Ureaconcentrationmmol/L

Ammonium ion productionafter aerobic preincubation(mmol/g faeces) [Median,Range]

Ammonium ion production afteraerobic preincubation (mmol/gfaeces) [Median, Range]

SIDS 1.91± 0.41 7.1± 11.0 [ 1.8, 0 - 39.5] 6.0± 11.1 [ 2.4, 0 - 46.7]SIDS 3.38± 0.47 8.8± 13.6 [ 1.3, 0 - 43.49] 8.2± 15.3 [ 3.1, 0 - 65.0]Controls 1.91± 0.41 24.9± 23.1 [ 19.3, 0 - 59.5] 23.8± 28.0 [ 9.4, 0.6 - 72.3] Controls 3.38± 0.47 32.5± 29.5 [ 24.4, 0 - 82.9] 29.9± 33.7 [ 12.8, 0.85 - 80.8]

Table 4. Mean sigmoid faecal ammonium ion production ± SD after 30h incubation with ureaas substrate at two different concentrations.

Faecal urea content and nitric oxide concentrations

The urea content of the faeces of SIDS infants was significantly greater than that of control

infants, (p<0.01) while there was no significant differences between the nitric oxide

concentrations. (p>0.05) (Table 5)

Faecal urea content (µmol/g) Nitric oxide concentration (ppb)SIDS 116 ± 75 492 ± 620Control 14 ± 16 233 ± 121

Table 5. Faecal urea content and nitric oxide concentrations ± SD of SIDS cases comparedwith control cases.

Paper III

Geographical and Seasonal Distribution of SIDS

In this paper the geographical and seasonal distribution of SIDS in relation to content of

nitrate in drinking water and groundwater levels were studied.

Incidence of SIDS and geographical and temporal distribution

The annual incidence of SIDS during 1990 through 1996 decreased from 1.30 to 0.46 deaths

per 1000 live births. Maximal incidence of SIDS occurred in the 13th week of life and the

median time of death was 11 weeks. The lowest incidence was 0.56 in the Västmanland

26

region and the highest was 1.15 per 1000 live births in the Västernorrland region. In 79 of the

total of 288 municipalities, which accounted for 11% of the total population, there were no

SIDS cases during the entire study period.

Seasonal incidence of SIDS and groundwater levels

Majority of SIDS deaths occurred between the months of September and April. Infants born

in a certain month were found to have a regional incidence as high as 7 - 8 times the mean

incidence for the whole country. The greatest incidence in the north of the country occurred at

a time when the south of the country had its lowest incidence. However this was not

statistically significant. SIDS cases occurred mostly at a time when groundwater was

increasing or high during spring thawing or winter rains. No cases occurred when frost

penetration and snow prevented groundwater discharge. The observed timing of deaths was

found to be related to high groundwater levels (p<0.001) and more so in small groundwater

basins (88% of deaths occurring at high groundwater levels) than in big basins (78% of deaths

at high groundwater levels).

Clusters

A statistically significant (p = 0.049) cluster of a high incidence of SIDS comprising 127

municipalities in central and north Sweden occurred during 1990 and 1991. A statistically

significant (p = 0.045) negative cluster comprising 33 adjacent municipalities in the counties

of Uppsala, Gävleborg, Södermanland, Västmanland and Dalarna occurred during 1995 and

1996. This negative cluster corresponded to a period with a decreasing national incidence as

well as low precipitation and low or decreasing groundwater levels in this part of the country.

Incidence of SIDS and the use of private wells

Amongst the SIDS cases, 16.8% had their water supply from private wells; this was similar to

that of the live births population, of whom 16.5% had their water supply from private wells.

The nitrate concentration (median) in the groundwater of the dug wells (private wells) was

27

found to be 10 - 20mg/L (range 0.5 - 70) and that of the drilled wells (public water supply)

was 1 - 5mg/L (range 0.5 - 70).

Incidence of SIDS and the nitrate content of drinking water from publicly owned waterworks

When the incidence of SIDS in a number of municipalities and its correlation to published

nitrate concentrations (medians) was studied using a Spearman rank correlation analysis, the

correlation coefficient was found to be low (n=53, rs=0.07, p>0.05); the correlation was

higher when a sample of public waterworks from 1989 which was available at SGU was used

(n=30, rs=0.74, p< 0.0001) or if the chemical content of the drinking water supply of a

number of municipalities obtained from records was used (n=18, rs=0.48, p=0.04). However,

the SIDS incidence was better correlated with published maximal nitrate concentrations

(n=53, rs=0.34, p=0.01) or if concentrations from a number of public waterworks were used

(n=18, rs=0.87, p=0.0002). The difference between the published maximal and median nitrate

concentrations and the corresponding incidence of SIDS was found to be correlated (n=53,

rs=0.51, p=0.0003); the correlation was greater using figures from the records of the public

waterworks (n=18, rs=0.92, p<0.0001). When the maximum concentration and the difference

between maximum and median concentrations were plotted against the corresponding

incidence of SIDS, the plot of the individual points showed a marked inflection point.

Paper IV

Carbon Dioxide Production

Carbon dioxide production in infants given 400 µmol/kg per day of ammonium chloride

orally was not significantly different from those of infants who did not receive any

supplements. Carbon dioxide production was found to be 0.62 (0.04 SE) mol/kg/d in infants

with ammonium supplementation and 0.59 (0.02 SE) mol/kg/d in infants without

supplementation. (p=0.35)

28

Total Energy Expenditure

The total energy expenditure values derived from the carbon dioxide production

measurements was 0.33 (0.2 SE) MJ/kg/d in infants given ammonium supplements and 0.32

(0.01 SE) MJ/kg/d in infants without supplementation, the difference was not significant.

Urine Urea, Ammonium Ion Concentration and pH

Urea, ammonium ion concentration and pH of the urine of the infants given ammonium

supplementation were 50.70 (7.32 SE) mmol/L, 5729 (1334 SE) µmol/L and 6.67 (0.15 SE)

respectively. The urea, ammonium ion concentration, and pH of the urine of infants without

ammonium supplementation were 49.32 (11.59 SE) mmol/L, 5500 (926 SE) µmol/L and 6.77

(0.17 SE) respectively. The differences between the infants with and without ammonium

supplementation were not significant and the p values calculated using the Wilcoxon signed

rank test for urea, ammonium ion concentration and pH were equal to 0.46, 0.17 and 0.25

respectively.

29

DISCUSSION

Urea Metabolism

Urea metabolism is important in the human infant and a derangement of this metabolism

could explain some of the factors known to be associated with SIDS. Our results in Paper II

show that faecal urease activity is impaired in SIDS resulting in significant amounts of

unmetabolized urea in faeces. The low faecal urease activity and the increased faecal urea

content differed significantly from that of infants who died of other causes; they are therefore

not merely the effects of postmortem decomposition.

Breastmilk is known to contain 3 - 6 mmol/L of urea (increasing in concentration during

infancy) and 100 - 500 µmol/L of ammonium ions (decreasing in concentration during

infancy). (3, 4) Ingested urea is completely metabolized by enteric microflora (mainly

Bifidobacteria, which are urease producing) in the gut of healthy infants. (Paper I) Urea

which is thus broken down to ammonium ions is absorbed from the gut and transported via

the splanchnic circulation to the liver. Some of the nitrogen (between 17 - 61%) salvaged

from the breakdown of urea in the gut is added to the nitrogen pool and may be utilised for

protein synthesis. (30, 38, 39) Rate of salvage of urea nitrogen is greatest during the first 6

weeks of life with a gradual slowing down thereafter. (13, 31) In the liver, ammonium is

metabolized to urea and carbon dioxide. (33) Urea - synthesizing capacity of the liver matures

in term and preterm infants during the first few weeks of life. (40, 41) In Paper IV addition of

400 µmol/kg/day of ammonium ions to the normal diet of breastmilk or formula feeds for 12

days did not augment carbon dioxide production in infants 2 to 6 months old. The dose of

ammonium ion given is equivalent to approximately two- and- a- half times the normal intake

in a healthy breastfed infant. A much larger oral dose of 2.8 mmol/kg body weight of

ammonium chloride is known to cause metabolic acidosis, (42) and this loading dose was

used by Adamovich et al to study the effects of metabolic acidosis on renal sodium

30

reabsorption in neonates. Healthy infants absorb almost 100% of the oral intake of ammonium

present in normal breasmilk or fomula feeds. (Paper I) Infants recovering from infections (8,

38) and low - birth weight infants (43) had greater rates of nitrogen retention than normal

infants (44). Compromised infants are thus capable of adapting to their increased nitrogen

demands. This could explain the results in our study on healthy infants (Paper IV)

Ammonium above the nitrogen needs of the healthy infant may have been lost in faeces or

urine. (45) However, urine urea and ammonium in infants given ammonium did not differ

significantly from those who were not given ammonium in our study. (Paper IV) Some of the

ammonium absorbed may have been incorporated into their body proteins. Heine et al found

that between 79 and 94% of labelled nitrogen instilled in the colon of infants, was retained in

their body protein pool. (11, 12)

Breastfeeding

Breastfeeding is known to have a protective effect against SIDS. (46, 47) Though modern

formula feeds have a low protein content and are enriched with lactose (8) to encourage the

growth of Bifidobacteria in the intestines, they were unable to suppress the growth of

Clostridia and Enterococci. (9) The urea concentration in formula feeds is equal to that of

cow`s milk (i.e. half of that of human breastmilk). In Paper I the faecal microflora of the

infants who were mostly breastfed were found to be predominantly Bifidobacteria though

aerobic bacteria such as Enterococci were present within 3 days of birth. A decrease in

Bifidobacteria because of an overwhelming increase in pathogenic bacteria or secondary to

antibiotic therapy could consequently decrease the breakdown of urea to ammonium ions. In

healthy infants the supply of urea and ammonium ions in formula feeds may be adequate.

However, in compromised infants, the breastmilk supply of urea and ammonium ions may be

critical.

31

Postmortem Vitreous Humor Levels of Urea Nitrogen

It is interesting to note that postmortem urea nitrogen levels in the vitreous humor of SIDS

infants were found to be significantly lower compared to those who had died of other causes.

(48) This is consistent with our findings of significant amounts of unmetabolized urea in the

faeces of SIDS infants. (Paper II) Postmortem vitreous humor chemistry reflects antemortem

blood chemistry and remains stable for up to 30 hours after death. (49)

Preceding Illnesses

Gastrointestinal illnesses such as vomiting and diarrhea, are reported by Hoffman (23) and

others (24) to occur more frequently in the preceding 2 weeks in SIDS cases than in living

control cases; this difference was not found by others. (50, 51) A higher proportion of

antibodies to clostridial and enterobacterial toxins and a lower proportion of maternal

antibodies were seen in SIDS infants compared to controls suggesting that SIDS infants

succumbed to these infections because of an inadequacy in their immune response to these

toxins. (52) A change in the normal microflora to a preponderance of pathogenic bacteria

could explain the significant amounts of unmetabolized faecal urea seen in SIDS infants.

However, we found that faecal microflora in infants who had succumbed to SIDS were

similar to those found in normal infants. (Paper I and II)

Helicobacter pylori Infection

Helicobacter pylori infection has been proposed as a possible cause of SIDS (53) and this has

been supported by others. (54) It has been suggested that urease produced by H. pylori if

aspirated by infants, in whom the gastrooesophageal reflux is commonly seen, could on

reaching the alveoli react with plasma urea, to produce ammonia. This ammonia,

hypothetically, could be then absorbed systemically producing symptoms of ammonia toxicity

resulting in respiratory arrest. (53) On the other hand, absorption of ammonia through the

gastric mucosa enables it to be detoxified by the liver. Both H. pylori and Bifidobacteria are

32

known to be urease producing and should be able to metabolize completely, the urea

contained in breastmilk and formula feeds. If SIDS is caused by a H. pylori infection, it does

not explain our finding of significant amounts of unmetabolized faecal urea in SIDS infants

compared to controls. (Paper II)

Carbon Dioxide Production and Total Energy Expenditure

In Paper IV no change in carbon dioxide production was found following ammonium

ingestion. The DLW technique was used to determine carbon dioxide production. As this is a

non - invasive technique not requiring the cooperation of the subjects, it is well suited for

investigations in infants. In a pilot study using the Datex monitor to measure carbon dioxide

production following ammonium chloride ingestion difficulties were encountered, as reliable

measurements were only possible in the resting or sleeping infant and this was not always the

case in the 90 minutes following ammonium ingestion.

Values of total energy expenditure (TEE) can be derived from carbon dioxide production

measurements determined by the DLW technique. In 1996 Butte et al (55) published a review

of available data of the TEE of infants, assessed by the DLW method, which revealed that

estimates of TEE during infancy tended to be lower when obtained by this method than from

estimates of dietary energy intake. An extensive data set published in 2000 (56) confirmed the

lower values of TEE, determined by DLW method compared to earlier values, which are used

in the recommendations for dietary energy by world health authorities such as the FAO. (55,

57) No such data has been published for Swedish infants. In Paper IV we produced these

data, which were found to be comparable to that produced by Butte et al, confirming that

present recommendations for dietary energy intake are in need of revision.

Smoking

Maternal smoking as a potential risk factor for SIDS has been reported as early as 1966 by

Steel et al, (58) and this has been confirmed by other researchers. (59, 60, 61) Passive smoke

33

exposure was also found to increase the incidence of SIDS, (62, 63) and a dose - response

relationship between a greater exposure to smoke and SIDS has been reported. (60, 63) Effect

of smoke on inflammatory mediators (64) and the presence of nicotine in breastmilk (65) may

contribute to the increased risk. Though several reasons have been postulated, no causal

mechanism has been elucidated. In Paper I we showed that nitric oxide, nitrate and nitrite

have an inhibitory effect on faecal urease, produced by enteric microflora. Nitric oxide

present in the ambient air of smokers, when inhaled by infants, could transform to nitrites

after dissolving in water, and then impair their enteral urease activity. Inhaled nitric oxide has

been demonstrated to be converted to nitrates and subsequently eliminated by the kidneys.

(66) Smoking mothers can be presumed to have increased nitrate levels in blood and

breastmilk. Inhaled nitric oxide and nitrates in breastmilk may together impair enteral urease

activity in the infant resulting in significant amounts of unmetabolized urea and consequently

decreasing the amounts of ammonium being transported to the infant liver. The protective

effect of breastfeeding was seen in nonsmokers and not in smokers. (63) Endogenous nitric

oxide production is known to increase during the neonatal period in normal term infants, and

then gradually decrease until the infant is 4 months old, when it reaches a plateau. Preterm

infants had greater nitric oxide productions than term infants. (67)

Prone Position and Body Temperature

Prone position, hyperthermia and overwrapping have been associated with an increased risk

of SIDS in several studies. (68, 69) Decreased ventilation as a result of heat stress has been

suggested as the common factor (70) leading to the final apnoea of SIDS. A newborn infant

with a low metabolic rate and a large surface area to mass ratio is prone to cold stress, while a

3 month old infant with a higher metabolic rate, lower surface area to mass ratio and better

tissue insulation is more prone to heat stress. (71, 72) Thus both cold and heat stress may

initiate or enhance hypoventilation especially in an infant with some degree of metabolic

alkalosis, which has been suggested to be present in SIDS infants. (34)

34

Increased Winter Incidence of SIDS

SIDS occurs most frequently in the winter months in both the northern and the southern

hemispheres. Infection, (73) overwrapping, and smoking in confined spaces have all been

previously suggested to contribute to the increased winter incidence. Peak levels of

environmental pollutants such as carbon monoxide, sulphur dioxide, nitrogen dioxide and

hydrocarbons have been shown to precede the increase of SIDS. (74) Animal studies have

demonstrated that nitrogen dioxide resulted in changes in the bronchial epithelium which

contributed to a susceptibility to infection. (75, 76) Prone sleeping partially accounted for the

increase in SIDS rates in winter but some unidentified seasonal factor was found to modify its

effect. (77)

In Paper III the seasonality of SIDS during 1990 through 1996 in Sweden was found to vary

according to the location of the infants` residences, with reference to the geographical

latitude. The incidence of SIDS was highest in the northernmost part of the country when it

was at its lowest in the rest of the country. The occurrence of individual deaths was associated

with the recharge of groundwater, which is known to increase its nitrate content. (78, 79, 80)

SIDS cases occurred when the groundwater level rose, mostly as a result of spring thawing or

autumn rains. No SIDS cases occurred in the northernmost parts of the country, when frost

penetration and snow prevented changes in the groundwater level. This is in agreement with a

study by Munro et al (81) who showed that infant mortality was more than 31 percent greater

in infants resident in waterlogged soils than on well drained soils. Change in the

environmental temperature rather than the exact temperature was found to be related to the

incidence of SIDS. (82) Greatest risk of SIDS was on the warmer days of winter when the

daytime average was 3oC greater than the preceding 30 days. (83) Thawing on the warmer

days following a colder spell with a resultant increase in water nitrate levels could be a

possible explanation.

35

CONCLUSION

1. Aerobic and anaerobic microflora appear in the faeces of healthy infants within 3 days of

birth and are present throughout the first 6 months of life. Faecal urease activity decreases

to a third while bacterial production of nitric oxide increases ten - fold during the same

period. Nitrate, nitrite and nitric oxide have an inhibitory effect on faecal urease activity.

2. An insufficient metabolism of enteric urea is associated with SIDS.

3. The seasonal distribution of SIDS varies widely in different parts of the country and is

associated with groundwater level changes. Peak or widely varying drinking water nitrate

concentrations is associated with the incidence of SIDS.

4. Ingestion of additional ammonium chloride, equivalent to at least two-and-a-half times the

normal amount found in breastmilk feeds, does not increase carbon dioxide production in

healthy infants.

Inadequate enteric urea and nitrogen metabolism in infants and peak nitrate levels in drinking

water are associated with the incidence of SIDS. Further evidence is required to demonstrate

if they play a causal role in SIDS.

Future Research

Several questions remain to be answered.

1. Do infants of smoking mothers have an impaired faecal urease activity and a high faecal

urea content compared to controls?

2. Are nitrate levels in drinking water in countries known to have a high incidence, greater

and more varying than those of countries with a low incidence?

3. Do infants at a high risk of SIDS have a metabolic alkalosis?

36

4. Does the addition of ammonium supplements to the feeds of low- weight infants and

infants with infections, increase their carbon dioxide productions?

37

ACKNOWLEDGEMENTS

This work was carried out at the Anaesthesiology and Intensive Care Unit, Department ofSurgical Sciences, Uppsala University Hospital, Sweden.

I wish to express my deepest gratitude to my mentor and supervisor Professor Lars Wiklund.His enthusism, expertise and generous help have been vital for the completion of this study. Isincerely appreciate his willingness at all times in promptly replying to my queries.

I am grateful to Elisabet Forsum, for sharing her knowledge and experience in the field ofstable isotopes, and as co- author.

My sincere thanks to Lars Holmberg, for unravelling the mysteries of epidemiology, and asco- author.

I wish to expres my gratitude to Gunnar Ronquist, Carl- Erik Nord, Andy Coward and AntonyWright, for analysing data, and as co- authors.

I wish to thank Göran Hedenstierna, Tom Saldeen, Björn Zarén, Lisa Wernroth, Jan Pousette,Mats Aastrup and Bo Thunholm for contributing their expertise to this study, and as co-authors.

My sincere appreciation for the Staff of the Paediatric Anaesthesia Section, The CentralIntensive Unit and the Anaesthetic Laboratory for their enthusiastic help and technicalsupport.

My deepest gratitude to the parents of all the infants who participated in the studies.

I wish to thank Siv Utterberg for invaluable administrative assistance.

My thanks to my colleagues at the Anaesthetic Department for their encouragement and foraccepting my absences from the clinical rota during my research weeks.

My heartfelt thanks to my friends, both far and near, for their enthusiastic support and genuineinterest in my scientific pursuits.

My warmest gratitude to my sister and brother and their respective families, for their lovingsupport. I also wish to thank my brother for finding solutions, to some of my computerproblems, at all hours of the day.

Financial support by the Uppsala University Hospital, the Marianne and Marcus WallenbergsFoundation and the Laerdal Foundation for Acute Medicine is gratefully acknowledged.

38

REFERENCES

1. Macy G. Composition of human colostrom and milk. Am J Dis Child 1949; 78: 589-595.

2. Lönnerdal B, Forsum E, Hambreus L. A longitudinal study of the protein, nitrogen, andlactose contents of human milk from Swedish well – nourished mothers. Am J Clin Nutr1976, 29. 1127 – 1133.

3. Hambreus L. Proprietary milk versus human breast milk in infant feeding. Pediatr ClinNorth Am 1977; 24: 17 – 36.

4. Harzer G, Franzke V, Bindels JG. Human milk nonprotein nitrogen components:changing patterns of free amino acids and urea in the course of early lactation. Am JClin Nutr 1984; 40. 303 – 309.

5. Bullen CL,Tearle PV, Willis AT. Bifidobacteria in the intestinal tract of infants: an in-vivo study. J Med Microbiol 1976, 9: 325 – 333.

6. Suzuki K, Benno Y, Mitsuoka T, Takebe S, Kobashi K, Hase J. Urease-producingspecies of intestinal anaerobes and their activities. Appl Environ Microbiol 1979; 37:379 – 382.

7. Crociani F, Matteuzzi D. Urease activity in the genus bifidobacterium. Ann Microbiol(Inst Pasteur) 1982,; 133A: 417 – 423.

8. Heine W, Tiess M, Stolpe HJ, Wutzke K. Urea utilization by the intestinal flora, ofinfants fed mother`s milk and a formula diet, as measured with the 15N technique. JPediatr Gastroenterol Nutr 1984; 3. 709 – 712.

9. Kleesen B, Bunke H, Tovar K, Noack J, Sawtzki. Influence of two infant formulas andhuman milk on the development of the faecal flora in newborn infants. Acta Pediatr1995; 84: 1347 – 1356.

10. Walser M, Bodenlos LJ. Urea metabolism in man. J Clin Invest 1959; 38: 1617 – 1625

11. Heine W, Wutzke KD, Richter I, Walther F, Plath c. Evidence for colonic absorption ofprotein nitrogen in infants. Acta Paediatr Scand 1987; 76: 741 - 744.

12. Heine W, Christa M, Wutzke KD. Host-microflora correlations in infant nutrition.Progress Food Nutr Sci 1992; 16: 181 – 197.

13. Fomon SJ, Matthews DE, Bier DM, Rogers RR, Rebouche CJ, Edwards BB, NelsonSE. Bioavailability of dietary urea nitrogen in the infant. J Pediatr 1987: 111: 221 –224.

14. Rodwell VW. Catabolism of proteins and of amino acids. In: Murray RK, Granner DK,Mayes PA, Rodwell VW, ed. Harper`s Biochemistry Appleton & Lange, 2000: 313 –322.

39

15. Valtin H, Schafer JA, ed. Role of kidneys in acid- base balance: Renal excretion of H+

and conservation of HCO3− . In: Renal Function Little, Brown and Co, 1995: 209 – 233.

16. Bourke E, Atkinson DE. The development of acid- base concepts and their applicationto mammalian pH homeostasis. Häussinger D, ed. In: pH Homeostasis. Mechanismsand Control Academic Press, 1988: 163 – 179.

17. The Bible. 1 Kings 3:19.

18. Valdes- Dapena M. A half century of progress: The evolution of SIDS research. In:Rognum TO, ed. Sudden Infant Death Syndrome. New Trends in the Nineties. Norway:Scandinavian University Press, 1995:3 –10.

19. Guilleminault C, Periata R, Souquet M, Dement WC. Apneas during sleep in infants:Possible relationship with sudden infant death syndrome. Science 1975; 190: 677 – 679.

20. Shannon DC, Kelly D. Impaired regulation of alveolar ventilation and the sudden infantdeath syndrome. Science 1977; 197: 367 – 368.

21. Naeye RL. Sudden infant death. Sci American 1980; 242: 52 –58.

22. Kinney HC, Filiano JJ, Harper RM. The neuropathology of the sudden infant deathsyndrome: A review. J Neuropathol Exp Neurol 1992; 51(2): 115 – 126.

23. Hoffman HJ, Damus K, Hillman L, Krongrad E. Risk factors for SIDS: The NationalInstitute of Child Health and Human Development SIDS Cooperative EpidemiologicStudy. Ann NY Acad Sci 1988; 533: 13 –30.

24. Hassall IB, Vendenberg J. Infant sleep position: A New Zealand survey. NZ Med J1985; 98: 97 – 99.

25. deJonge GA, Engelberts AC, Koomen – Liefting AJM, Kostense PJ. Cot death andprone sleeping position in the Netherlands. Br Med J 1989; 298: 722 – 724.

26. Beal SM, Finch CF. An overview of retrospective case – control studies investigatingthe relationship between prone sleeping position and SIDS. J Pediatr Child Health1991, 27: 334 –339.

27. Wigfield RE, Fleming PJ, Berry PJ, Rudd PT, Golding J. Can the fall in Avon`s suddeninfant death rate be explained by changes in sleeping position? Br Med J 1992; 304: 282– 283.

28. Lundequist B, Nord CE, Winberg J. The composition of the faecal microflora inbreastfed and bottle fed infants from birth to eight weeks. Acta Paediatr Scand 1985;74: 45 – 51.

29. Heine W, Mohr C, Wutzke KD, Radke M. Symbiotic interactions between colonicmicroflora and protein metabolism in infants. Acta Paediatr Scand 1991; 80. 7 –12.

40

30. Wheeler RA, Jackson AA, Griffiths DM. Urea production and recycling in neonates. JPediatr Surg 1991; 26: 575 – 575.

31. Steinbrecher HA, Griffiths DM, Jackson AA. Urea production in normal breast-fedinfants measured with primed/intermittent oral doses of [15N, 15N] urea. Acta Paediatr1996; 85: 656 – 662.

32. Jackson AA. Urea as a nutrient: bioavailability and role in nitrogen economy. Arch DisChild 1994, 70: 3 – 4.

33. Atkinson DE, Camien MN. The role of urea synthesis in the removal of metabolicbicarbonate and the regulation of pH. Curr Top Cell Regul 1982; 21: 261 – 302.

34. Wiklund L. Carbon dioxide formation and elimination in man. Recent theories andpossible consequences. Upsala J Med Sci 1996; 101: 35 – 68.

35. Kulldorff M. A spatial span statistic. Communications in Statistics: Theory and Methods1997; 26: 1481 - 1496.

36. Ritz P, Coward WA. Doubly labelled water measurement of total energy expenditure.Diabete Metab 1995, 21: 241 –251.

37. Prentice AM, Diaz EO, Murgatroyd PR, Goldberg GR, Sonko BJ, Black AE, CowardWA. Doubly Labeled Water Measurements In Practice. In: Whitehead RG & Cole TJ,ed. New Techniques in Nutritional Research. Academic Press 1991: 177 –200.

38. Heine W, Tiess M, Wutzke KD. 15N tracer investigations of the physiologicalavailability of dietary urea nitrogen in mother`s milk. Acta Paediatr Scand 1986; 75:439 – 443.

39. Jackson AA. Optimizing amino acid and protein supply and utilization in the newborn.Proc Nutr Soc 1989; 48: 293 - 301.

40. Boehm G, Teichmann B, Jung K. Development of urea-synthesizing capacity in preterminfants during the first weeks of life. Biol Neonate 1991; 59: 1 – 4.

41. Boehm G, Gedlu E, Mueller MD, Beyreiss K, Räihä NCR. Postnatal development ofurea– and ammonia- excretion in urine of very-low-birth-weight infants small forgestational age. Acta Paediatrica Hungarica 1991; 31(1): 31 – 45

42. Adamovich K, Sulyok E, Guignard JP. Effects of NH4Cl Metabolic acidosis on

proximal and distal sodium reabsorption in newborn infants. Biol Neonate 1991; 60:193 - 200.

43. Donovan SM, Lönnerdal B, Atkinson SA. Bioavailability of urea nitrogen for the lowbirthweight infant. Acta Paediatr Scand 1990; 79: 899 – 905.

44. Fomon SJ, Bier DM, Mathews DE, et al. Bioavailability of dietary urea nitrogen in thebreast- fed infant. J Pediatr 1988; 113:5115 – 517.

41

45. Darling P, Wykes LJ, Clarke R, Papageorgiou, Pencharz. Utilization of non- proteinnitrogen in whey- dominant formulae by low- birthweight infants. Clin Sci 1993; 84:543 – 548.

46. Mitchell EA, Scragg R, Stewart AW, Becroft DMO,Ford RPK, et al. Results from thefirst year of the New Zealand cot death study. NZ Med J 1991; 104:71 – 76.

47. Hoffman HJ, Hillman LS. Epidemiology of the sudden infant death syndrome:maternal, neonatal, and postneonatal risk factors. Clinics Perinat 1992; 19: 717 – 737.

48. Blumenfeld TA, Mantell CH, Ctherman RL, Blanc WA. Postmortem vitreous humorchemistry in sudden infant death syndrome and in other causes of death in childhood.Am J Clin Pathol 1979; 71: 219 – 223.

49. Swift PGF, Worthy E, Emery JL. Biochemical state of the vitreous humour of infants atnecropsy. Arch Dis Child 1974; 49: 680 – 685.

50. Gilbert RE, Fleming PJ Azaz Y, Rudd PT. Signs of illness preceding suddenunexpected death in infants. BMJ 1990; 300: 1237 – 1239.

51. Ponsonby AL, Dwyer T, Kasl SV, Cochrane JA. The Tasmanian SIDS case- controlstudy: univariable and multivariable risk factor analysis. Paediatr Perinat Epid 1995; 9:256 – 272.

52. Siarakas S, Brown AJ, Murrell WG. Immunological evidence for a bacterial toxinaetiology in sudden infant death syndrome. FEMS Immunol Med Micrbiol 1999; 25: 37– 50.

53. Pattison CP, Marshall BJ. Proposed link between Helicobacter pylori and sudden infantdeath syndrome. Med Hypothesis 1997; 49: 365 – 369.

54. Kerr JR, Al- Khattal A, Barson AJ, Burnie JP. An association between sudden infantdeath syndrome (SIDS) and Helicobacter pylori infection. Arch Dis Child 2000; 83: 429– 434.

55. Butte NF. Energy requirements of infants. Eur J Clin Nutr 1996, 50: Suppl. 1. S24 -

S36.

56. Butte NF, William WW, Hopkinson JM, Mehta NR, Smith EO. Energy expenditurederived from energy expenditure and energy deposition during the first 2y of life. Am JClin Nutr 2000; 72: 1558 – 1569.

57. Butte NF, Henry CJK, Torun B. Report of the working group on energy requirements ofinfants, children and adolescents. Eur J Clin Nutr 1996; 50: Suppl. 1. S188 - S189.

58. Steele R, Langworth JT. The relationship of antenatal and postnatal factors to suddenunexpected death in infancy. Canad Med Ass J 1966; 94: 1165 – 1171.

59. Mitchell EA, Ford RPK, Stewart AW, Taylor BJ, Becroft DMO, Thompson JMD, et al.Smoking and the sudden infant death syndrome. Pediatrics 1993; 91: 393 – 396.

42

60. Haglund B, Cnattingius S. Cigarette smoking as a risk factor for sudden infant deathsyndrome: a population- based study. Am J Public Health 1990; 80: 29 – 32.

61. Haglund B. Cigarette smoking and sudden infant death syndrome: some salient points inthe debate. Acta Paediatr Suppl 1993; 389: 37 – 39.

62. Scragg R, Mitchell EA, Taylor BJ, Stewart AW, Ford RPK, et al. Bed sharing, smoking,and alcohol in the sudden infant death syndrome. Br Med J 1993; 307: 1312 – 1318.

63. Klonoff- Cohen HS, Edelstein SL, Lefkowitz ES, Srinivasan IP, Kaegi D, et al. Theeffect of passive smoking and tobacco exposure through breast milk on sudden infantdeath syndrome. JAMA 1995; 273: 795 – 798.

64. Raza MW, Essery SD, Elton RA, Weir DM, Busuttil A et al. Exposure to cigarettesmoke, a major risk factor for sudden infant death syndrome: effects of cigarette smokeon inflammatory responses to viral infection and bacterial toxins. FEMS Immunologyand Medical Microbiology 1999; 25: 145 – 154.

65. Rylander E, Pershagen G, Eriksson M, Berman G. Parental smoking, urinary cotinine,and wheezing bronchitis in children. Epidemiology 1995; 6: 289 – 293.

66. Westfelt UN, Benthin G, Lundin S, Stenquist O, Wennmalm Å. Conversion of inhalednitric oxide to nitrate in man. Br J Pharmacol 1995; 114: 1621 – 1624.

67. Tsukahara H, Hiraoka M, Hori C, Tsuchida S, Hata I, et al. Urinary nitrite/ nitrateexcretion in infancy: comparison between term and preterm infants. Early Human Dev1997; 47: 51 – 56.

68. Fleming PJ, Gilbert R, Azaz Y, Berry PJ, Rudd PT et al. Interactions between beddingand sleeping position in sudden infant death syndrome: a population based case- controlstudy. Br Med J 1990; 301: 85 – 89.

69. Ponsonby A- L, Dwyer T, Gibbons LE, Cochrane JA, Jones M et al. Thermalenvironment and sudden infant death syndrome: case – control study. Br Med J. 1992;304: 277 – 282.

70. Gozal D, Colin AA, Daskalovic YI Jaffe M. Environmental overheating as a cause oftransient respiratory chemoreceptor dysfunction in an infant. Pediatrics 1988; 82: 738 –740.

71. Wailoo MP, PetersenSA, Whittaker H, Goodenough P. The thermal environment inwhich 3- 4 month old infants sleep at home. Arch Dis Child 1989; 64: 600 – 604.

72. Fleming PJ, Levine MR, Azaz Y, Wigfield R, Stewart AJ. Interactions betweenthermoregulation and the control of respiration in infants: possible relationship tosudden infant death. In: Lagercrantz H, Tunell R, Wennergren G, ed. Proceedings of aSwedish Medical Research Council State of the Art Conferene on the Sudden infantdeath syndrome. Acta Paediatr 1993; Suppl 389: 57 – 59.

43

73. Helweg – Larsen K, Lundemose JB, Oyen N, Skjaerven R, Alm B, et al. Interactions ofinfectious symptoms and modifiable risk factors in sudden infant death syndrome. TheNordic Epidemiological SIDS study. Acta Paediatr 1999; 88: 521 – 527.

74. Hoppenbrouwers T, Calub M, Arakawa K, Hodgman J. Seasonal relationship of suddeninfant death syndrome and environmental pollutants. Am J Epidemiol 1981; 113: 623 –635.

75. Ehrlich R. Effect of nitrogen dioxide on resistance to respiratory infection. Bact Rev1966; 30:604 – 614.

76. Stephens RJ, Freeman G,Crane SC, ET AL. Ultrastructural changes in the terminalbronchiole of the rat during continuous, low- level exposure to nitrogen dioxide. ExpMol Pathol 1971; 14: 1 – 19.

77. Mitchell EA, Clements M, Williams SM, Stewart AW, Cheng A. Seasonal differencesin risk factors for sudden infant death sydrome. Acta Paediatr 1999; 88: 253 – 258.

78. Creed IF, Band LE, Foster NW, Morrison IK, Nicolson JA. Regulation of nitratre – Nrelease from temperate forests: A test of the N flushing hypothesis. Water ResourcesResearch 1996; 32: 3337 – 3354.

79. Liu ZJ, Hallberg GR, Malanson GP. Structural equation modeling of dynamics ofnitrate contamination in ground water. J Am Water Resour Assoc 1997; 33:1219 – 1235.

80. Arheimer B, Andersson L, Lepistö A. Variation of nitrogen concentration in foreststreams – influences of flow, seasonality and catchment characteristics. J Hydrology1996; 179: 281 – 304.

81. Munro LJA, Penning – Rowsell EC, Barnes HR, Fordham MH, Jarrett D. Infantmortality and soil type: a case study in south – central England (with discussion). Eur JSoil Sci 1997; 48: 1 – 17.

82. Campbell MJ. Sudden infant death sndrome and environmental temperature: furtherevidence for a larger relationship. Med J Aust 1989; 151: 365 – 367.

83. Schluter PJ, Ford RPK, Brown J, Ryan AP. Weather temperatures and sudden deathsyndrome: a regional study over 22 years in New Zealand. J Epidemol CommunityHealth 1998; 52: 27 – 33.