1170
Better Perinatal Health SWEDEN GÖSTA ROOTH Perinatal Research Unit,
Department of Pædiatrics, University of Uppsala, Sweden
WHY is Swedish perinatal mortality so low? A short is that perinatal mortality rates close to ours (9-4 1000 in 1978) can be found in other places where per there is similar awareness of the problem and the mothers are equally healthy. answer
DATA COLLECTION
Since 1749, Sweden and Finland have collected nationwide
figures for births and deaths, from which table I on infant mortality is derived. Then as now, parish priests kept the primary records; most people lived in the country; and by law everyone had to go to church. Therefore the early data are probably reliable. Over the past 230 years there has been a gradual reduction in the infant mortality. The accompanying figure shows the data from 1915 to 1977, including the two components of perinatal mortality, late fetal deaths and early neonatal deaths. By about 1940 they were both falling and they are now the same at 4.77 per 1000 each. Awareness of perinatal mortality became stronger in the 1930s. Sweden then had few hospitals for delivery and few children’s hospitals, the latter mainly in the largest cities, Stockholm, Goteborg and Maimo, as well as at the universities of Uppsala and Lund. Babies were born either at home or (again mainly in the largest cities) in small private delivery units often located in ordinary apartment houses. As late as 1940 these private units were regarded as the best place to have a child for those who could afford it. During the depression in the 1930s, the birth rate fell noticeably, and maternal and child care became a major political issue, advanced particularly by Alva and Gunnar Myrdal.1 A parliamentary commission produced a series of reports which led to complete remodelling of the system. They called for free antenatal care for all pregnant women, the building of hospitals with delivery and pxdiatric units, and free health care for all newborn infants. These recommendations were based partly on the association between poverty and high infant mortality as revealed by Rietz2 (table II): the commission stated that every pregnant woman should have access by right to the sort of care hitherto available only to the well-off.3 Thus, among other benefits, perinatal mortality would fall and the effects of the low birth rate would be partly offset. TABLE I-SWEDISH INFANT MORTALITY PER
1751-1976
1000 LIVEBORN,
Infant
mortality
and
perinatal mortality in Sweden, 1915-77.
The third major stage of awareness stemmed from the introduction of "medical birth registration". In 1969, on the initiative of Prof. Peter Karlberg,4 the National Board of Health and Social Welfare began to study the feasibility of a common nationwide system for recording details of antenatal care, delivery, and care of the infant. Medical birth registration-a summary of the main events during pregnancy and delivery, and after the birth-was introduced rather abruptly in 1973 because Sweden had been selected to participate in a WHO study, Social and Biological Effects on Perinatal Mortality.5
The main data from 1973 and 1974 are published6 and a report covering 1975 and 1976 will appear soon. Detailed information is thus to hand, but reference to social factors is almost completely lacking because the computer staff are not allowed to incorporate such information. Printouts are available and one important function is to provide every delivery unit in Sweden with its own perinatal statistics three times a year, with data from other units for comparison. These figures are being watched closely by doctors, hospital administrators, and politicians. Medical birth registration is also used for special studies such as the influence of paracervical blockade on perinatal mortality and the clustering of specific birth defects. Table III is taken from preliminary printouts for 1978. AVAILABILITY OF SERVICES
Despite
the clamour for home
fanatics who
are
delivery from a few getting their views
very successful in
into the news media, virtually all deliveries in Sweden take place in hospitals. In 1977 there were only 4 planned home deliveries. There were 141 deliveries in women awaiting transport or in transit to hospital. Oddly, this happened about as often in Stockholm as in the north of Sweden with its great distances. Table v shows perinatal mortality in relation to the type of hospital. For smaller units the mean value of 87 TABLE II-INFANT MORTALITY AMONG LEGITIMATE INFANTS BORN
1918-22
IN STOCKHOLM IN RELATION TO THE INCOME OF
THE
*
from Rietz.2
FAMILY*
1171 TABLE III-DATA FROM PRELIMINARY COMPUTER PRINTOUT FROM MEDICAL BIRTH REGISTRATION
1978
composite of both high and low values. Ideally the perinatal mortality in such units should be zero because they are supposed to refer complications early to specialised departments. The persistent perinatal mortality in some of these units shows that, even with 99% attendance at antenatal clinics, as there is all over Sweden, the uncomplicated cases cannot be selected with certainty; indeed, medical birth registration revealed that about half the total perinatal mortality in Sweden occurred in is
a
previously uncomplicated pregnancies.
The numbers are and the of the small delivery small, apparent advantage in close to has been units, home, interpreted other ways: have made use of this emotional issue, often politicians the figures indiscriminately. Nevertheless the citing trend is distinct. In 1950, 1973, and 1978 45, 70, and 74% of the total deliveries were conducted in the well equipped, larger obstetric units. Maternity homes disappeared in 1966. Since perinatal mortality is much influenced by birthweight, the data in table Iv may be more useful for comparison than overall perinatal mortality. PREVENTION, DIAGNOSIS,
MANAGEMENT
The
virtually ubiquitous antenatal and postnatal clinics, mainly for preventive care, are a characteristic of Swedish medicine. Run by the local district doctor with a midwife or nurse, or in the cities usually by specialists, these clinics are within easy reach of almost everyone. Thus, Uppsala with about 100 000 inhabitants has 10 clinics and also central referral clinics at the University Hospital. As a rule the nurses get to know TABLE V-MEDICAL BIRTH REGISTRATION
1978:
the local population very well and the acceptance of these clinics is such that about 99% of pregnant women attend for all the planned visits from the beginning of pregnancy until the infant is 4 years old. There is no reward, no compulsion; it is just done. In the twelve antenatal visits which are customary, a doctor sees the mother twice. During the first year of life the infant is seen four to six times by the doctor and the nurse, and may in addition visit the clinical health nurse any time the mother feels the need. An important point is that the health nurse visits every single newborn infant at home within the first week after return from the hospital. As well as keeping standard records of events in and after pregnancy the clinics provide uniform recommendations on what to do under certain well-defined circumstances. These nationwide medical recommendations were considerably improved in 1973, at the time when medical birth registration was introduced; their effect is that all doctors work according to one pattern. This is an important factor in the persistent decline in Swedish perinatal mortality. Regionalisation is the systematic referral, within a given part of the country, of complicated cases to the central or regional specialised hospital; only now being recognised as important in the U.S.A., this has been working for many years in Sweden. Table v also gives information about the number of psediatric departments. More relevant in this connection is the number of intensive-care units. The first was established in 1973. Although so far we have no formal neonatal intensive-care units in Sweden, five university hospitals have a total of some 30 beds where artificial respiration may be given and ten hospitals can assemble about 100 beds for special care of the newborn. A measure of the good overall health of our newborn infants is that, in a year, each unit sees only about 20 infants needing artificial ventilation. As to asphyxia, respiratory distress, and so on, despite the uniform protocols and the WHO definitions there is so much variation in diagnosis that no useful figure can be offered. TABLE IV-PERINATAL MORTALITY IN DIFFERENT WEIGHT GROUPS IN
1978
(PRELIMINARY DATA)
NUMBER OF BIRTHS AND PERINATAL MORTALITY FOR DIFFERENT TYPES OF HOSPITALS
1172 TRENDS IN PERINATAL CARE
Obstetricians are taking over more and more of the antenatal care. They also supervise an increasing number of labours and deliveries, though they still leave normal delivery to the midwives. An increasingly important part of the prenatal medical service is to acquaint the mother and the father with the delivery hospital, its facilities, and the various techniques it offers. The Swedish Parliament in 1971 stated tht every woman in labour has the right to modern analgesia, but allocated no money for this. 100% nitrous oxide was the routine analgesia until 15 years ago; today it is 50/50 nitrous oxide/oxygen and pudendal blockade. Paracervical blockade has been, and still is, powerfully advocated in some areas, but more and more epidurals are being done. The number of epidurals is limited chiefly by the shortage of anaesthetists. The steady drop in perinatal mortality over a long period-and even that during the five years 1973-78, when the rate fell from 14-3 to 9-4 per 1000-cannot be attributed to any single factor. Continuous fetalheart-rate monitoring can be done in all the larger delivery units, but it is used variably. Fetal scalp blood sampling is far too rarely practised in too few places; probably only one or two of the university clinics do it regularly, and the exceptional central hospital. Ultrasound measurements are rapidly being introduced, but on the whole are still a novelty. Westin’s approach of identifying (inter alia) twins and small-for-dates fetuses by regular measurement of the symphysis-fundus distance is being universally adopted in Sweden.7 This seems particularly relevant since, malformations apart, smallness for dates is the main cause of fetal and neonatal deaths. It may seem that this presentation has dealt too much with obstetrics and too little with pxdiatrics. But if healthy infants are delivered no paediatric care is needed, and if preventive medicine is effective the obstetricians, too, have less work to do. Everyone is aware of the need to "dehospitalise" childbirth. Today the father is welcomed at all normal deliveries, and often also at caesarean section; the rooms
becoming more homelike, though delivery rooms are on pictures and flowers. Rooming-in is generally accepted, and so are visits by older children. Parents are encouraged to attend their infants in the intensiveare
still short
unit. Will our the largest
care
1000,
so
perinatal mortality continue to fall? Some of hospitals now claim figures of about 8 per with more general application of existing
methods we can expect the overall rate to fall to that level at least. Such further reduction may seem small, but each 1 per 1000 means 100 infants saved in a year. Moreover, two or three times that number will be spared birth injuries. In antenatal care we must improve the recognition of risk groups such as twins, the offspring of diabetics, and small-for-dates infants, so the best care can start early. At hospital level the need is for closer interaction between obstetricians, neonatologists, and research-workers. Most of the data for this paper
were
kindly supplied by
Mr Anders
Ericsson, National Board of Health and Welfare. He and Dr Lars-Eric made valuable suggestions. A series of pertinent recent be found in Perspectives of Child Health in Sweden, Acta Pædiat Scand suppl 275, 1979.
Bratteby
papers is
to
Separate reprints of this article are not available, reprint of the series will be published later.
but
a
collected
REFERENCES
Myrdal A, Myrdal G. Kns i befolkningsfrågan (Crisis in population growth). Stockholm: Bonniers, 1934. 2. Rietz E. Sterblichkeit und Todesursachen in den Kinderjahren. Acta Pædiat Scand suppl 3, 1930. 1.
3. Slutbetänkande.
Avgivet av Befolkningskommissionen. Utredningar. Stockholm, 1938: 57.
Statens
Offentliga
Karberg P, Priolisi A. Clinical analysis of causes of death with emphasis on perinatal mortality. Monogr Pædiat no. 9, 1977. 5. World Health Organisation. Social and biological effects on perinatal mortality. Vol I-II, Geneva: WHO, 1978. 6. Medical Birth Registration in 1973 and 1974. Statistics of the National Board of Health and Social Welfare. Statistiska Centralbyrån, Liber distribution, S-162 89 Vällingby, Sweden, 1976. 7. Westin B. Gravidogram and fetal growths: Comparison with biochemical supervision. Acta Obstet Gynæc Scand 1977; 56: 273-82.
4.
Occasional
Survey
IDENTIFICATION OF SITE OF URINARY-TRACT INFECTIONS BY ANTIBODY-COATED BACTERIA ASSAY KENNETH A. MUNDT
B. FRANK POLK
Channing Laboratory and Department of Medicine, Peter Hospital Division of the Affiliated Hospitals
Bent Brigham
Center, Inc., Boston, Massachusetts, U.S.A. Accurate identification of the site of urinary-tract infection (UTI) may be clinically and epidemiologically valuable. A review of the literature on the assessment of the antibody-coated bacteria (ACB) assay, a non-invasive technique introduced five years ago, shows that, compared with acceptable standards (bilateral ureteral catheterisation or bladder washout), the overall sensitivity of the ACB assay is 83·1%, the specificity is 76·7%, the predictive value positive is 81·3%, and the predictive value negative is 78·8%. These findings suggest that the ACB assay has at present no role in the management of patients with UTI. Its usefulness as an epidemiological tool remains to be demonstrated.
Summary
INTRODUCTION
DETERMINATION of the site of urinary-tract infection (UTI) has been not only a challenge to clinicians and investigators but also a point of controversy. Some investigators believe that knowledge of the site of infection makes for successful therapy,’.2 and others claim that identification of renal involvement will allow those affected to be followed up more closely;3 however, Kunin suggests that localisation is of little clinical value (although he concedes its usefulness as a tool for the study of the pathogenesis and natural history of urinarytract infections).4 Data. relevant to the natural history of symptomatic urinary-tract infections are both scanty and conflicting, a simple reliable method of identifying the site of infection is needed.5 None of the diagnostic methods is entirely satisfactory.6 Methods of identifying site of infection may be either invasive ("direct") or non-invasive ("indirect"). In general, the invasive methods--e.g., bilateral ureteral
Better Perinatal Health SWEDEN GÖSTA ROOTH Perinatal Research Unit,
Department of Pædiatrics, University of Uppsala, Sweden
WHY is Swedish perinatal mortality so low? A short is that perinatal mortality rates close to ours (9-4 1000 in 1978) can be found in other places where per there is similar awareness of the problem and the mothers are equally healthy. answer
DATA COLLECTION
Since 1749, Sweden and Finland have collected nationwide
figures for births and deaths, from which table I on infant mortality is derived. Then as now, parish priests kept the primary records; most people lived in the country; and by law everyone had to go to church. Therefore the early data are probably reliable. Over the past 230 years there has been a gradual reduction in the infant mortality. The accompanying figure shows the data from 1915 to 1977, including the two components of perinatal mortality, late fetal deaths and early neonatal deaths. By about 1940 they were both falling and they are now the same at 4.77 per 1000 each. Awareness of perinatal mortality became stronger in the 1930s. Sweden then had few hospitals for delivery and few children’s hospitals, the latter mainly in the largest cities, Stockholm, Goteborg and Maimo, as well as at the universities of Uppsala and Lund. Babies were born either at home or (again mainly in the largest cities) in small private delivery units often located in ordinary apartment houses. As late as 1940 these private units were regarded as the best place to have a child for those who could afford it. During the depression in the 1930s, the birth rate fell noticeably, and maternal and child care became a major political issue, advanced particularly by Alva and Gunnar Myrdal.1 A parliamentary commission produced a series of reports which led to complete remodelling of the system. They called for free antenatal care for all pregnant women, the building of hospitals with delivery and pxdiatric units, and free health care for all newborn infants. These recommendations were based partly on the association between poverty and high infant mortality as revealed by Rietz2 (table II): the commission stated that every pregnant woman should have access by right to the sort of care hitherto available only to the well-off.3 Thus, among other benefits, perinatal mortality would fall and the effects of the low birth rate would be partly offset. TABLE I-SWEDISH INFANT MORTALITY PER
1751-1976
1000 LIVEBORN,
Infant
mortality
and
perinatal mortality in Sweden, 1915-77.
The third major stage of awareness stemmed from the introduction of "medical birth registration". In 1969, on the initiative of Prof. Peter Karlberg,4 the National Board of Health and Social Welfare began to study the feasibility of a common nationwide system for recording details of antenatal care, delivery, and care of the infant. Medical birth registration-a summary of the main events during pregnancy and delivery, and after the birth-was introduced rather abruptly in 1973 because Sweden had been selected to participate in a WHO study, Social and Biological Effects on Perinatal Mortality.5
The main data from 1973 and 1974 are published6 and a report covering 1975 and 1976 will appear soon. Detailed information is thus to hand, but reference to social factors is almost completely lacking because the computer staff are not allowed to incorporate such information. Printouts are available and one important function is to provide every delivery unit in Sweden with its own perinatal statistics three times a year, with data from other units for comparison. These figures are being watched closely by doctors, hospital administrators, and politicians. Medical birth registration is also used for special studies such as the influence of paracervical blockade on perinatal mortality and the clustering of specific birth defects. Table III is taken from preliminary printouts for 1978. AVAILABILITY OF SERVICES
Despite
the clamour for home
fanatics who
are
delivery from a few getting their views
very successful in
into the news media, virtually all deliveries in Sweden take place in hospitals. In 1977 there were only 4 planned home deliveries. There were 141 deliveries in women awaiting transport or in transit to hospital. Oddly, this happened about as often in Stockholm as in the north of Sweden with its great distances. Table v shows perinatal mortality in relation to the type of hospital. For smaller units the mean value of 87 TABLE II-INFANT MORTALITY AMONG LEGITIMATE INFANTS BORN
1918-22
IN STOCKHOLM IN RELATION TO THE INCOME OF
THE
*
from Rietz.2
FAMILY*
1171 TABLE III-DATA FROM PRELIMINARY COMPUTER PRINTOUT FROM MEDICAL BIRTH REGISTRATION
1978
composite of both high and low values. Ideally the perinatal mortality in such units should be zero because they are supposed to refer complications early to specialised departments. The persistent perinatal mortality in some of these units shows that, even with 99% attendance at antenatal clinics, as there is all over Sweden, the uncomplicated cases cannot be selected with certainty; indeed, medical birth registration revealed that about half the total perinatal mortality in Sweden occurred in is
a
previously uncomplicated pregnancies.
The numbers are and the of the small delivery small, apparent advantage in close to has been units, home, interpreted other ways: have made use of this emotional issue, often politicians the figures indiscriminately. Nevertheless the citing trend is distinct. In 1950, 1973, and 1978 45, 70, and 74% of the total deliveries were conducted in the well equipped, larger obstetric units. Maternity homes disappeared in 1966. Since perinatal mortality is much influenced by birthweight, the data in table Iv may be more useful for comparison than overall perinatal mortality. PREVENTION, DIAGNOSIS,
MANAGEMENT
The
virtually ubiquitous antenatal and postnatal clinics, mainly for preventive care, are a characteristic of Swedish medicine. Run by the local district doctor with a midwife or nurse, or in the cities usually by specialists, these clinics are within easy reach of almost everyone. Thus, Uppsala with about 100 000 inhabitants has 10 clinics and also central referral clinics at the University Hospital. As a rule the nurses get to know TABLE V-MEDICAL BIRTH REGISTRATION
1978:
the local population very well and the acceptance of these clinics is such that about 99% of pregnant women attend for all the planned visits from the beginning of pregnancy until the infant is 4 years old. There is no reward, no compulsion; it is just done. In the twelve antenatal visits which are customary, a doctor sees the mother twice. During the first year of life the infant is seen four to six times by the doctor and the nurse, and may in addition visit the clinical health nurse any time the mother feels the need. An important point is that the health nurse visits every single newborn infant at home within the first week after return from the hospital. As well as keeping standard records of events in and after pregnancy the clinics provide uniform recommendations on what to do under certain well-defined circumstances. These nationwide medical recommendations were considerably improved in 1973, at the time when medical birth registration was introduced; their effect is that all doctors work according to one pattern. This is an important factor in the persistent decline in Swedish perinatal mortality. Regionalisation is the systematic referral, within a given part of the country, of complicated cases to the central or regional specialised hospital; only now being recognised as important in the U.S.A., this has been working for many years in Sweden. Table v also gives information about the number of psediatric departments. More relevant in this connection is the number of intensive-care units. The first was established in 1973. Although so far we have no formal neonatal intensive-care units in Sweden, five university hospitals have a total of some 30 beds where artificial respiration may be given and ten hospitals can assemble about 100 beds for special care of the newborn. A measure of the good overall health of our newborn infants is that, in a year, each unit sees only about 20 infants needing artificial ventilation. As to asphyxia, respiratory distress, and so on, despite the uniform protocols and the WHO definitions there is so much variation in diagnosis that no useful figure can be offered. TABLE IV-PERINATAL MORTALITY IN DIFFERENT WEIGHT GROUPS IN
1978
(PRELIMINARY DATA)
NUMBER OF BIRTHS AND PERINATAL MORTALITY FOR DIFFERENT TYPES OF HOSPITALS
1172 TRENDS IN PERINATAL CARE
Obstetricians are taking over more and more of the antenatal care. They also supervise an increasing number of labours and deliveries, though they still leave normal delivery to the midwives. An increasingly important part of the prenatal medical service is to acquaint the mother and the father with the delivery hospital, its facilities, and the various techniques it offers. The Swedish Parliament in 1971 stated tht every woman in labour has the right to modern analgesia, but allocated no money for this. 100% nitrous oxide was the routine analgesia until 15 years ago; today it is 50/50 nitrous oxide/oxygen and pudendal blockade. Paracervical blockade has been, and still is, powerfully advocated in some areas, but more and more epidurals are being done. The number of epidurals is limited chiefly by the shortage of anaesthetists. The steady drop in perinatal mortality over a long period-and even that during the five years 1973-78, when the rate fell from 14-3 to 9-4 per 1000-cannot be attributed to any single factor. Continuous fetalheart-rate monitoring can be done in all the larger delivery units, but it is used variably. Fetal scalp blood sampling is far too rarely practised in too few places; probably only one or two of the university clinics do it regularly, and the exceptional central hospital. Ultrasound measurements are rapidly being introduced, but on the whole are still a novelty. Westin’s approach of identifying (inter alia) twins and small-for-dates fetuses by regular measurement of the symphysis-fundus distance is being universally adopted in Sweden.7 This seems particularly relevant since, malformations apart, smallness for dates is the main cause of fetal and neonatal deaths. It may seem that this presentation has dealt too much with obstetrics and too little with pxdiatrics. But if healthy infants are delivered no paediatric care is needed, and if preventive medicine is effective the obstetricians, too, have less work to do. Everyone is aware of the need to "dehospitalise" childbirth. Today the father is welcomed at all normal deliveries, and often also at caesarean section; the rooms
becoming more homelike, though delivery rooms are on pictures and flowers. Rooming-in is generally accepted, and so are visits by older children. Parents are encouraged to attend their infants in the intensiveare
still short
unit. Will our the largest
care
1000,
so
perinatal mortality continue to fall? Some of hospitals now claim figures of about 8 per with more general application of existing
methods we can expect the overall rate to fall to that level at least. Such further reduction may seem small, but each 1 per 1000 means 100 infants saved in a year. Moreover, two or three times that number will be spared birth injuries. In antenatal care we must improve the recognition of risk groups such as twins, the offspring of diabetics, and small-for-dates infants, so the best care can start early. At hospital level the need is for closer interaction between obstetricians, neonatologists, and research-workers. Most of the data for this paper
were
kindly supplied by
Mr Anders
Ericsson, National Board of Health and Welfare. He and Dr Lars-Eric made valuable suggestions. A series of pertinent recent be found in Perspectives of Child Health in Sweden, Acta Pædiat Scand suppl 275, 1979.
Bratteby
papers is
to
Separate reprints of this article are not available, reprint of the series will be published later.
but
a
collected
REFERENCES
Myrdal A, Myrdal G. Kns i befolkningsfrågan (Crisis in population growth). Stockholm: Bonniers, 1934. 2. Rietz E. Sterblichkeit und Todesursachen in den Kinderjahren. Acta Pædiat Scand suppl 3, 1930. 1.
3. Slutbetänkande.
Avgivet av Befolkningskommissionen. Utredningar. Stockholm, 1938: 57.
Statens
Offentliga
Karberg P, Priolisi A. Clinical analysis of causes of death with emphasis on perinatal mortality. Monogr Pædiat no. 9, 1977. 5. World Health Organisation. Social and biological effects on perinatal mortality. Vol I-II, Geneva: WHO, 1978. 6. Medical Birth Registration in 1973 and 1974. Statistics of the National Board of Health and Social Welfare. Statistiska Centralbyrån, Liber distribution, S-162 89 Vällingby, Sweden, 1976. 7. Westin B. Gravidogram and fetal growths: Comparison with biochemical supervision. Acta Obstet Gynæc Scand 1977; 56: 273-82.
4.
Occasional
Survey
IDENTIFICATION OF SITE OF URINARY-TRACT INFECTIONS BY ANTIBODY-COATED BACTERIA ASSAY KENNETH A. MUNDT
B. FRANK POLK
Channing Laboratory and Department of Medicine, Peter Hospital Division of the Affiliated Hospitals
Bent Brigham
Center, Inc., Boston, Massachusetts, U.S.A. Accurate identification of the site of urinary-tract infection (UTI) may be clinically and epidemiologically valuable. A review of the literature on the assessment of the antibody-coated bacteria (ACB) assay, a non-invasive technique introduced five years ago, shows that, compared with acceptable standards (bilateral ureteral catheterisation or bladder washout), the overall sensitivity of the ACB assay is 83·1%, the specificity is 76·7%, the predictive value positive is 81·3%, and the predictive value negative is 78·8%. These findings suggest that the ACB assay has at present no role in the management of patients with UTI. Its usefulness as an epidemiological tool remains to be demonstrated.
Summary
INTRODUCTION
DETERMINATION of the site of urinary-tract infection (UTI) has been not only a challenge to clinicians and investigators but also a point of controversy. Some investigators believe that knowledge of the site of infection makes for successful therapy,’.2 and others claim that identification of renal involvement will allow those affected to be followed up more closely;3 however, Kunin suggests that localisation is of little clinical value (although he concedes its usefulness as a tool for the study of the pathogenesis and natural history of urinarytract infections).4 Data. relevant to the natural history of symptomatic urinary-tract infections are both scanty and conflicting, a simple reliable method of identifying the site of infection is needed.5 None of the diagnostic methods is entirely satisfactory.6 Methods of identifying site of infection may be either invasive ("direct") or non-invasive ("indirect"). In general, the invasive methods--e.g., bilateral ureteral
