Saturday 3 January 1976
INTRACEREBRAL HÆMORRHAGE Accuracy of Computerised Transverse Axial Scanning in Predicting the Underlying Aetiology R. D. HAYWARD
G. V. A. O’REILLY
Departments of Neurosurgery and Neuroradiology, Atkinson Morleys Hospital, 31 Copse Hill, London SW20 Summary
The ised
diagnostic transverse
accuracy of computeraxial tomography in
the pathology underlying cases of nontraumatic intracerebral hæmatoma has been reviewed. A 160 × 160 matrix was used and the scans of one hundred patients whose diagnoses had been confirmed by either angiographic or post-mortem examination were studied. In the absence of any clinical information a diagnosis of either cerebral aneurysm or primary intracerebral hæmorrhage could be made with an accuracy of 90%. The hæmatomas most likely to be misinterpreted are those associated with angiomas and tumours, and these two lesions were responsible for six out of the ten incorrect diagnoses. In predicting the site of a ruptured aneurysm accuracy ranged from 100%, for those involving the anterior cerebral artery complex to 66% for those few hæmatomas associated with aneurysms of the internal carotid/proximal middle cerebral artery aneurysm complex.
predicting
Introduction THIS study was undertaken to demonstrate a significant change in our ability to discover the setiology underlying non-traumatic intracerebral hxmorrhage without recourse to angiography. The size of this clinical problem can be seen from table i which gives the numbers of cerebral aneurysms, angiomas, and primary intracerebral hxmorrhages seen at this hospital over the past three years. In the past, the diagnosis of cerebral aneurysm has been made angiographically, while primary intracerebral haemorrhage has been diagnosed by exclusion-that is, angiograms have failed to show an underlying aneurysm, or TABLE I-PATHOLOGY OF NON-TRAUMATIC INTRACRANIAL HAEMORRHAGE SEEN AT THIS HOSPITAL
presence of hxmatoma. Paxton and Ambrosel first demonstrated the ability of computerised transverse axial (C.T.A.) scanning to detect intracerebral blood-cot, which has a higher photon absorption coefficient than brain and appears as a well demarcated "white" area. As familiaritywith this investigation has grown, it has become apparent that in many cases it is possible not only to predict the -presence of an underlying aneurysm, but also to forecast the site of origin of that aneurysm.
angioma in the
Cromptonzhas pointed out that published informaregard to the anatomy of intracerebral haematoma was sparse. His series was based solely on postmortem findings, but we have been able to demonstrate those anatomical sites whose involvement may help a clinician to make an accurate diagnosis in the living pation with
tient. Methods Routine scanning was not performed on all patients presenting with non-traumatic intracanial haemorrhage. The scans described below were performed on patients selected for the following reasons:
(1) Before angiography, where known or suspected haemorrhage associated with signs of hemisphere dysfunction (hemiparesis, dysphasia, visual-field disturbance, etc.) (2) Following angiography for intracranial haemorrhage when a previously unsuspected hasmatoma was seen. (3) When angiography in intracranial hxmorrhage had shown two or more possible lesions (aneurysms, angiomas etc.) and the site of a htmatoma on the scan would indicate the lesion responsible for the haemorrhage.
was
scanned in the transverse axial plane 160 matrix. The scans were displayed on a cathode-ray tube and ’Instamatic’ ’Polaroid’ exposures made. The scans of all patients with non-traumatic intracerebral
All
using
patients
a
160
were
x
hsematomas from February, 1974 (the introduction of the 160 x 160 matrix), to October, 1975, were studied. Patients under twenty-one years of age and those whose heematomas were in the posterior fossa were excluded. The scans were viewed by both authors together without recourse to any clinical information (such as name, age, blood-pressure, and clinical his-
tory). After the
scans
had been viewed and their
recorded, the patients’
interpretation
reviewed. Those who had neither angiographic nor post-mortem confirmation of the C.T.A. scan findings were excluded from the study. Also excluded were those patients whose scans were considered to be unsuitable for study on technical grounds. notes were
Results A total of one hundred and twelve scans were reviewed of which ten were rejected because there was 7949
2 TABLE
II-C.T.A.-BASED PREDICTION OF PATHOLOGY UNDERLYING NON-TRAUMATIC INTRACEREBRAL IIaeMATOMA I
I
I
I
being situated in the temporal lobe, due to a ruptured middle cerebral artery aneurysm. The scans in which the presence of an aneurysm had been predicted were now divided into four further groups
(1) Aneurysms of the anterior cerebral complex (including communicating artery). (2) Aneurysms at the main branchings of the middle cerebral the anterior
artery.
neither angiographic nor post-mortem confirmation for the predictions made. Of the remaining one hundred and two scans, only one depended upon postmortem examination alone without previous angiography to confirm the c.T.. scan findings. A further two scans were rejected because their quality did not allow a reasonable prediction to be made, one being due to patient movement aberration and the other because it did not include cuts low enough to evaluate the possible origin of the hxmatoma. An eventual total of one hundred scans was selected for this study. They were first divided into four possible groups (aneurysms, primary intracerebral haematoma, angioma, or tumour) depending on the suspected underlying pathology. The results of this study are given in table 11. A diagnosis of aneurysm or primary intracerebral hxmatoma can be made with approximately 9000 accuracy (table n). However, when the figures are viewed in the light of the true pathology as demontrated by angiography or necropsy it can be seen that three out of a possible total of thirty-eight primary intracerebral hxmatomas were missed. Of these three, two had been diagnosed as middle cerebral aneurysms because of the presence of hasmatoma in the region of the Sylvian fissure. One hxmatoma in the occipital region had been incorrectly diagnosed as an angioma. Only one aneurysm was missed out of a possible total of fifty-two. This was a large aneurysm situated at the carotid termination which had produced a haematoma splitting upwards into the head of the caudate nucleus. Table 11 shows that, although four out of five angiomas were correctly predicted, a further four were diagnosed incorrectly. There were two tumours in the series, both of which had presented clinically as cases of intracerebral haemorrhage, and neither were recognised from their scans. One was thought to be a primary intracerebral haemorrhage and the other,
(3) Aneurysms of the internal carotid artery complex (posterior communicating artery aneurysms, terminal carotid artery aneurysms, and aneurysms of the proximal middle cerebral artery). (4) Other aneurysms. The results of these predictions are given in table ill. The accuracy varies from 100% for aneurysms of the anterior cerebral complex to 66% for those few hxmatomas associated with aneurysms of the internal carotid complex. The haematomas were further TABLE III-ACCURACY OF C.T.A. SCAN PREDICTIONS FOR EACH ANEURYSM SITE
*Posterior cerebral artery aneurysm.
show the anatomical sites of the various brain areas involved. Table iv shows how the presence of hasmatoma in the cavum of the septum pellucidum or corpus callosum can be taken as an absolute indication of an anterior cerebral complex aneurysm. A middle cerebral artery aneurysm can usually be identified if the associated haematoma follows the curve of the Sylvian fissure. All haematomas in the internal capsule were due to a primary intracerebral haemorrhage. Illustrations of typical intracerebral haematomas are given in figs 1-6.
analysed
to
Discussion This study demonstrates how, even in the absence of clinical information, it is possible to predict the aetiology underlying a non-traumatic intracerebral hxmatoma with an accuracy of 90%. Previous studies of primary intracerebral hxmatoma3 4 have depended
TABLE IV-C.T.A. SCAN AREAS INVOLVED BY NON-TRAUMATIC INTRACEREBRAL HaeMATOMA
*Haematoma following the
curve
of the
Sylvian fissure.
tHaematoma present in the basal cisterns.
3
Fig. 1-Primary hæmatoma in head of left caudate nucleus. Fig. 2—Haematoma from anterior cerebral artery aneurysm
following curve
of corpus callosum.
Fig. 3-Typical "comma-shaped" haematoma from ruptured middle cerebral artery aneurysm.
percutaneous carotid angiography to exclude cases of aneurysm and angioma. As this procedure can carry a significant morbidity-rate5 the emergence of a non-invasive scanning technique with this degree of accuracy is obviously important. The majority of these scans were made in patients in whom the presence of an intracerebral haematoma was suspected on the basis of clinical history, lumbar puncture, etc., prior to angiography, and the addition of the relevant clinical information will further improve the pre-angiographic diagnostic accuracy. In many cases of primary intracerebral haemorrhage it should be possible to dispense with angiography altogether. Conversely when a cerebral aneurysm is predicted from the position of its associated hxmatoma, the angiogram may be delayed until it is required for specific technical reasons-e.g., to provide information about arterial spasm, aneurysm size, or disposition of adjacent cerebral vessels. The incidence of hxmatoma formation at different intracerebral sites due to aneurysm is similar to Crompton’s2 post-mortem series reported from this hospital. In primary intracerebral hxmorrhage he found the external capsule most frequently involved (50% of cases) followed by the thalamus (18%). on
Fig. 4-Hoematoma in septum of cavum pellucidum from an anterior communicating artery aneurysm. Fig. 5-Large primary haematoma splitting laterally from region of external capsule. Fig. 6-Primary haematoma splitting along external capsule.
in the head of the caudate nucleus was but absolute indication of primary This contrasts with our series of live haemorrhage. patients where only twelve out of thirty-eight primary haematomas involved the external capsuale, eleven the thalamus, and eleven the head of the caudate nucleus. These figures imply that hxmorrhages involving the head of the caudate nucleus do not often cause death, presumably because of their small size. A further difference we have noted from Crompton’s series is in the frequency with which aneurysms of the middle cerebral artery produce hxmatomas involving the external capsule. We found this region to be an uncommon site for haemorrhage associated with an aneurysm but very common for primary hxmorrhage. The opposite was true for the para-sylvian region which was a common site for middle cerebral aneurysm hxmatomas, but uncommon for primary haemorrhage. Only two out of twenty middle cerebral artery aneurysms had haematomas affecting the external capsule (compared with Crompton’s eight out of sixteen), and only two primary intracerebral hxmorrhages involved the para-sylvian region. In Crompton’s post-mortem series haemorrhage from a middle
Haemorrhage said
to
be
a rare
4 cerebral artery aneurysm in the temporal lobe (six out of thirty-four cases) was comparatively rare compared with our incidence of seventeen hxmatomas in twenty cases. Presumably all these differences reflect the selection of patients into ante-mortem and postmortem groups by the severity of their haemorrhage. We thank Dr J. Ambrose, director of neuroradiology, and the consultant neurosurgeons of Atkinson Morley’s Hospital for allowing us to study the investigations of patients under their care. 1. 2. 3. 4. 5.
REFERENCES Paxton, R., Ambrose, J. Br. J. Radiol. 1974, 47, 530. Crompton, M. R. J. Neurol. Neurosurg. Psychiat. 1962, 25, 378. McKissock, W., Richardson, A. E., Walsh, L. S. Lancet, 1959, ii, 683. McKissock, W., Richardson, A. E., Taylor, J. ibid. 1961, ii, 221. Perrett, G., Nishoka, H. J. Neurosurg. 1966, 25, 98.
ULTRASOUND AND HUMAN-PLACENTAL-LACTOGEN SCREENING FOR EARLY DETECTION OF TWIN PREGNANCIES P.-H. PERSSON L. GRENNERT S. KULLANDER G. GENNSER
Department of Obstetrics and Gynœcology, University of Lund, Allmanna Sjukhuset, S-214 01 Malmö, Sweden J. THORELL Department of Nuclear Medicine, University of Lund, Allmänna Sjukhuset, Malmö In the past five years gradual introduction of ultrasonic screening of pregnant women increased the ante-partum detection of multiple pregnancy from 60% in 1971 to 95% in the first half of 1975. At the same time the average gestational age for detection of multiple pregnancies diminished from thirty-three to twenty-five weeks. The corresponding figures for 1963 were 32% and thirty-six weeks. Plasma humanplacental-lactogen concentrations were assessed for their value in selecting a smaller target group for subsequent ultrasonic screening. All but 2 of 39 twin pregnancies examined by single H.P.L. determinations had H.P.L. values more than 1 S.D. above the mean of the normal distribution. The use of plasma-H.P.L. screening might lower the proportion of patients requiring ultrasonography for antepartum diagnosis of multiple pregnancies to 16% of the total pregnant population.
from 20% to of suspicion multiple pregnancy can be confirmed by technical procedures, the crucial point being how to select the patients for these tests. We have evaluated the feasibility of screening an entire pregnant
figures reported varying delivery, 9 10 A clinical
49%.4
population for multiple pregnancies by ultrasonography in the second trimester. In a search for a way of lowering the number of ultrasonic investigations needed, the value of plasma human placental lactogen (H.P.L.) as a
selecting procedure was examined. Patients and Methods The investigation was done over eighteen months at Allmanna Sjukhuset, the sole maternity unit in the city of Matmo serving 250 000 inhabitants. At the first antenatal medical examination within the first four months of gestation, the women were referred for ultrasonic screening. During the first half of the period of investigation, only women attending the free antenatal clinics were invited and ultrasonic screening was performed between the twenty-fourth and thirtieth gestational weeks.
extended
to
During the second period, the investigation was practically all pregnant women within the city,
and the examinations were done between the twentieth and twenty-fifth weeks. The women were examined with a Kretz-Technic 4100 MGS or a ’Combison II’ echoscope with scan converter by either of two doctors or two specially trained midwives. Plasma-H.P.L. concentrations" were determined on bloodsamples taken at the ultrasonic screening. The coefficient of variation was 2.5-3% intra-assay and not more than 6% interassay. In one group of 25 women with twin pregnancies detected by ultrasound plasma-H.P.L. was measured serially during the pregnancy.
Summary
Results
During the first twelve months all 28 twin pregnancies of 2333 (about 80% of the pregnant women in the district) submitted for ultrasonic screening were diagnosed. During the last six months 1588 pregnant women were examined (95% of the total), and of the 15 twin pregnancies screened during this period 14 were correctly diagnosed. The diagnostic efforts increased the ante-partum detection of multiple pregnancy from 60% in 1971 to 95% in the first half of 1975 (fig. 1). In the past five years the average gestational age for detection of multiple pregnancies fell from thirty-three to twentyfive weeks with the gradual introduction of ultrasonic out
Introduction
Twin
pregnancies are associated with a high incidence of pre-term deliveries,l-3 and immaturity is a major factor in the neonatal mortality of twins.1 4 In 1952 Bender5
suggested that the pregnant woman should be to hospital to try to prevent pre-term delivery. Although the value of bed rest for the purpose of increasing the duration of gestation has been doubted,6-8 most investigators agree that the intrauterine growth of twins is encouraged by this approach.’8 despite differing opinions on the best measures for prevention of prematurity, the early diagnosis of twin pregnancy is accepted as a prerequisite for adequate antenatal care and management of delivery.34 7A surprisingly large number of twin pregnancies remain undetected until admitted
0[] UNKNOWN
INTRA-PARTUM
Fm []
X-RAY FOR OTHER PURPOSE
ULTRASONIC SCREENING
El CLINICAL Fig. 1—Method
of detection of twin pregnancies at Allinanna Sjukhuset, Maimo, before and after introduction of ultrasound.
INTRACEREBRAL HÆMORRHAGE Accuracy of Computerised Transverse Axial Scanning in Predicting the Underlying Aetiology R. D. HAYWARD
G. V. A. O’REILLY
Departments of Neurosurgery and Neuroradiology, Atkinson Morleys Hospital, 31 Copse Hill, London SW20 Summary
The ised
diagnostic transverse
accuracy of computeraxial tomography in
the pathology underlying cases of nontraumatic intracerebral hæmatoma has been reviewed. A 160 × 160 matrix was used and the scans of one hundred patients whose diagnoses had been confirmed by either angiographic or post-mortem examination were studied. In the absence of any clinical information a diagnosis of either cerebral aneurysm or primary intracerebral hæmorrhage could be made with an accuracy of 90%. The hæmatomas most likely to be misinterpreted are those associated with angiomas and tumours, and these two lesions were responsible for six out of the ten incorrect diagnoses. In predicting the site of a ruptured aneurysm accuracy ranged from 100%, for those involving the anterior cerebral artery complex to 66% for those few hæmatomas associated with aneurysms of the internal carotid/proximal middle cerebral artery aneurysm complex.
predicting
Introduction THIS study was undertaken to demonstrate a significant change in our ability to discover the setiology underlying non-traumatic intracerebral hxmorrhage without recourse to angiography. The size of this clinical problem can be seen from table i which gives the numbers of cerebral aneurysms, angiomas, and primary intracerebral hxmorrhages seen at this hospital over the past three years. In the past, the diagnosis of cerebral aneurysm has been made angiographically, while primary intracerebral haemorrhage has been diagnosed by exclusion-that is, angiograms have failed to show an underlying aneurysm, or TABLE I-PATHOLOGY OF NON-TRAUMATIC INTRACRANIAL HAEMORRHAGE SEEN AT THIS HOSPITAL
presence of hxmatoma. Paxton and Ambrosel first demonstrated the ability of computerised transverse axial (C.T.A.) scanning to detect intracerebral blood-cot, which has a higher photon absorption coefficient than brain and appears as a well demarcated "white" area. As familiaritywith this investigation has grown, it has become apparent that in many cases it is possible not only to predict the -presence of an underlying aneurysm, but also to forecast the site of origin of that aneurysm.
angioma in the
Cromptonzhas pointed out that published informaregard to the anatomy of intracerebral haematoma was sparse. His series was based solely on postmortem findings, but we have been able to demonstrate those anatomical sites whose involvement may help a clinician to make an accurate diagnosis in the living pation with
tient. Methods Routine scanning was not performed on all patients presenting with non-traumatic intracanial haemorrhage. The scans described below were performed on patients selected for the following reasons:
(1) Before angiography, where known or suspected haemorrhage associated with signs of hemisphere dysfunction (hemiparesis, dysphasia, visual-field disturbance, etc.) (2) Following angiography for intracranial haemorrhage when a previously unsuspected hasmatoma was seen. (3) When angiography in intracranial hxmorrhage had shown two or more possible lesions (aneurysms, angiomas etc.) and the site of a htmatoma on the scan would indicate the lesion responsible for the haemorrhage.
was
scanned in the transverse axial plane 160 matrix. The scans were displayed on a cathode-ray tube and ’Instamatic’ ’Polaroid’ exposures made. The scans of all patients with non-traumatic intracerebral
All
using
patients
a
160
were
x
hsematomas from February, 1974 (the introduction of the 160 x 160 matrix), to October, 1975, were studied. Patients under twenty-one years of age and those whose heematomas were in the posterior fossa were excluded. The scans were viewed by both authors together without recourse to any clinical information (such as name, age, blood-pressure, and clinical his-
tory). After the
scans
had been viewed and their
recorded, the patients’
interpretation
reviewed. Those who had neither angiographic nor post-mortem confirmation of the C.T.A. scan findings were excluded from the study. Also excluded were those patients whose scans were considered to be unsuitable for study on technical grounds. notes were
Results A total of one hundred and twelve scans were reviewed of which ten were rejected because there was 7949
2 TABLE
II-C.T.A.-BASED PREDICTION OF PATHOLOGY UNDERLYING NON-TRAUMATIC INTRACEREBRAL IIaeMATOMA I
I
I
I
being situated in the temporal lobe, due to a ruptured middle cerebral artery aneurysm. The scans in which the presence of an aneurysm had been predicted were now divided into four further groups
(1) Aneurysms of the anterior cerebral complex (including communicating artery). (2) Aneurysms at the main branchings of the middle cerebral the anterior
artery.
neither angiographic nor post-mortem confirmation for the predictions made. Of the remaining one hundred and two scans, only one depended upon postmortem examination alone without previous angiography to confirm the c.T.. scan findings. A further two scans were rejected because their quality did not allow a reasonable prediction to be made, one being due to patient movement aberration and the other because it did not include cuts low enough to evaluate the possible origin of the hxmatoma. An eventual total of one hundred scans was selected for this study. They were first divided into four possible groups (aneurysms, primary intracerebral haematoma, angioma, or tumour) depending on the suspected underlying pathology. The results of this study are given in table 11. A diagnosis of aneurysm or primary intracerebral hxmatoma can be made with approximately 9000 accuracy (table n). However, when the figures are viewed in the light of the true pathology as demontrated by angiography or necropsy it can be seen that three out of a possible total of thirty-eight primary intracerebral hxmatomas were missed. Of these three, two had been diagnosed as middle cerebral aneurysms because of the presence of hasmatoma in the region of the Sylvian fissure. One hxmatoma in the occipital region had been incorrectly diagnosed as an angioma. Only one aneurysm was missed out of a possible total of fifty-two. This was a large aneurysm situated at the carotid termination which had produced a haematoma splitting upwards into the head of the caudate nucleus. Table 11 shows that, although four out of five angiomas were correctly predicted, a further four were diagnosed incorrectly. There were two tumours in the series, both of which had presented clinically as cases of intracerebral haemorrhage, and neither were recognised from their scans. One was thought to be a primary intracerebral haemorrhage and the other,
(3) Aneurysms of the internal carotid artery complex (posterior communicating artery aneurysms, terminal carotid artery aneurysms, and aneurysms of the proximal middle cerebral artery). (4) Other aneurysms. The results of these predictions are given in table ill. The accuracy varies from 100% for aneurysms of the anterior cerebral complex to 66% for those few hxmatomas associated with aneurysms of the internal carotid complex. The haematomas were further TABLE III-ACCURACY OF C.T.A. SCAN PREDICTIONS FOR EACH ANEURYSM SITE
*Posterior cerebral artery aneurysm.
show the anatomical sites of the various brain areas involved. Table iv shows how the presence of hasmatoma in the cavum of the septum pellucidum or corpus callosum can be taken as an absolute indication of an anterior cerebral complex aneurysm. A middle cerebral artery aneurysm can usually be identified if the associated haematoma follows the curve of the Sylvian fissure. All haematomas in the internal capsule were due to a primary intracerebral haemorrhage. Illustrations of typical intracerebral haematomas are given in figs 1-6.
analysed
to
Discussion This study demonstrates how, even in the absence of clinical information, it is possible to predict the aetiology underlying a non-traumatic intracerebral hxmatoma with an accuracy of 90%. Previous studies of primary intracerebral hxmatoma3 4 have depended
TABLE IV-C.T.A. SCAN AREAS INVOLVED BY NON-TRAUMATIC INTRACEREBRAL HaeMATOMA
*Haematoma following the
curve
of the
Sylvian fissure.
tHaematoma present in the basal cisterns.
3
Fig. 1-Primary hæmatoma in head of left caudate nucleus. Fig. 2—Haematoma from anterior cerebral artery aneurysm
following curve
of corpus callosum.
Fig. 3-Typical "comma-shaped" haematoma from ruptured middle cerebral artery aneurysm.
percutaneous carotid angiography to exclude cases of aneurysm and angioma. As this procedure can carry a significant morbidity-rate5 the emergence of a non-invasive scanning technique with this degree of accuracy is obviously important. The majority of these scans were made in patients in whom the presence of an intracerebral haematoma was suspected on the basis of clinical history, lumbar puncture, etc., prior to angiography, and the addition of the relevant clinical information will further improve the pre-angiographic diagnostic accuracy. In many cases of primary intracerebral haemorrhage it should be possible to dispense with angiography altogether. Conversely when a cerebral aneurysm is predicted from the position of its associated hxmatoma, the angiogram may be delayed until it is required for specific technical reasons-e.g., to provide information about arterial spasm, aneurysm size, or disposition of adjacent cerebral vessels. The incidence of hxmatoma formation at different intracerebral sites due to aneurysm is similar to Crompton’s2 post-mortem series reported from this hospital. In primary intracerebral hxmorrhage he found the external capsule most frequently involved (50% of cases) followed by the thalamus (18%). on
Fig. 4-Hoematoma in septum of cavum pellucidum from an anterior communicating artery aneurysm. Fig. 5-Large primary haematoma splitting laterally from region of external capsule. Fig. 6-Primary haematoma splitting along external capsule.
in the head of the caudate nucleus was but absolute indication of primary This contrasts with our series of live haemorrhage. patients where only twelve out of thirty-eight primary haematomas involved the external capsuale, eleven the thalamus, and eleven the head of the caudate nucleus. These figures imply that hxmorrhages involving the head of the caudate nucleus do not often cause death, presumably because of their small size. A further difference we have noted from Crompton’s series is in the frequency with which aneurysms of the middle cerebral artery produce hxmatomas involving the external capsule. We found this region to be an uncommon site for haemorrhage associated with an aneurysm but very common for primary hxmorrhage. The opposite was true for the para-sylvian region which was a common site for middle cerebral aneurysm hxmatomas, but uncommon for primary haemorrhage. Only two out of twenty middle cerebral artery aneurysms had haematomas affecting the external capsule (compared with Crompton’s eight out of sixteen), and only two primary intracerebral hxmorrhages involved the para-sylvian region. In Crompton’s post-mortem series haemorrhage from a middle
Haemorrhage said
to
be
a rare
4 cerebral artery aneurysm in the temporal lobe (six out of thirty-four cases) was comparatively rare compared with our incidence of seventeen hxmatomas in twenty cases. Presumably all these differences reflect the selection of patients into ante-mortem and postmortem groups by the severity of their haemorrhage. We thank Dr J. Ambrose, director of neuroradiology, and the consultant neurosurgeons of Atkinson Morley’s Hospital for allowing us to study the investigations of patients under their care. 1. 2. 3. 4. 5.
REFERENCES Paxton, R., Ambrose, J. Br. J. Radiol. 1974, 47, 530. Crompton, M. R. J. Neurol. Neurosurg. Psychiat. 1962, 25, 378. McKissock, W., Richardson, A. E., Walsh, L. S. Lancet, 1959, ii, 683. McKissock, W., Richardson, A. E., Taylor, J. ibid. 1961, ii, 221. Perrett, G., Nishoka, H. J. Neurosurg. 1966, 25, 98.
ULTRASOUND AND HUMAN-PLACENTAL-LACTOGEN SCREENING FOR EARLY DETECTION OF TWIN PREGNANCIES P.-H. PERSSON L. GRENNERT S. KULLANDER G. GENNSER
Department of Obstetrics and Gynœcology, University of Lund, Allmanna Sjukhuset, S-214 01 Malmö, Sweden J. THORELL Department of Nuclear Medicine, University of Lund, Allmänna Sjukhuset, Malmö In the past five years gradual introduction of ultrasonic screening of pregnant women increased the ante-partum detection of multiple pregnancy from 60% in 1971 to 95% in the first half of 1975. At the same time the average gestational age for detection of multiple pregnancies diminished from thirty-three to twenty-five weeks. The corresponding figures for 1963 were 32% and thirty-six weeks. Plasma humanplacental-lactogen concentrations were assessed for their value in selecting a smaller target group for subsequent ultrasonic screening. All but 2 of 39 twin pregnancies examined by single H.P.L. determinations had H.P.L. values more than 1 S.D. above the mean of the normal distribution. The use of plasma-H.P.L. screening might lower the proportion of patients requiring ultrasonography for antepartum diagnosis of multiple pregnancies to 16% of the total pregnant population.
from 20% to of suspicion multiple pregnancy can be confirmed by technical procedures, the crucial point being how to select the patients for these tests. We have evaluated the feasibility of screening an entire pregnant
figures reported varying delivery, 9 10 A clinical
49%.4
population for multiple pregnancies by ultrasonography in the second trimester. In a search for a way of lowering the number of ultrasonic investigations needed, the value of plasma human placental lactogen (H.P.L.) as a
selecting procedure was examined. Patients and Methods The investigation was done over eighteen months at Allmanna Sjukhuset, the sole maternity unit in the city of Matmo serving 250 000 inhabitants. At the first antenatal medical examination within the first four months of gestation, the women were referred for ultrasonic screening. During the first half of the period of investigation, only women attending the free antenatal clinics were invited and ultrasonic screening was performed between the twenty-fourth and thirtieth gestational weeks.
extended
to
During the second period, the investigation was practically all pregnant women within the city,
and the examinations were done between the twentieth and twenty-fifth weeks. The women were examined with a Kretz-Technic 4100 MGS or a ’Combison II’ echoscope with scan converter by either of two doctors or two specially trained midwives. Plasma-H.P.L. concentrations" were determined on bloodsamples taken at the ultrasonic screening. The coefficient of variation was 2.5-3% intra-assay and not more than 6% interassay. In one group of 25 women with twin pregnancies detected by ultrasound plasma-H.P.L. was measured serially during the pregnancy.
Summary
Results
During the first twelve months all 28 twin pregnancies of 2333 (about 80% of the pregnant women in the district) submitted for ultrasonic screening were diagnosed. During the last six months 1588 pregnant women were examined (95% of the total), and of the 15 twin pregnancies screened during this period 14 were correctly diagnosed. The diagnostic efforts increased the ante-partum detection of multiple pregnancy from 60% in 1971 to 95% in the first half of 1975 (fig. 1). In the past five years the average gestational age for detection of multiple pregnancies fell from thirty-three to twentyfive weeks with the gradual introduction of ultrasonic out
Introduction
Twin
pregnancies are associated with a high incidence of pre-term deliveries,l-3 and immaturity is a major factor in the neonatal mortality of twins.1 4 In 1952 Bender5
suggested that the pregnant woman should be to hospital to try to prevent pre-term delivery. Although the value of bed rest for the purpose of increasing the duration of gestation has been doubted,6-8 most investigators agree that the intrauterine growth of twins is encouraged by this approach.’8 despite differing opinions on the best measures for prevention of prematurity, the early diagnosis of twin pregnancy is accepted as a prerequisite for adequate antenatal care and management of delivery.34 7A surprisingly large number of twin pregnancies remain undetected until admitted
0[] UNKNOWN
INTRA-PARTUM
Fm []
X-RAY FOR OTHER PURPOSE
ULTRASONIC SCREENING
El CLINICAL Fig. 1—Method
of detection of twin pregnancies at Allinanna Sjukhuset, Maimo, before and after introduction of ultrasound.
