Minor Physical Anomalies and Obstetric Complications in Schizophrenia

Abstract

Objective

To evaluate the possibility of using congenital minor physical anomalies (MPA) and obstetric complications (OC) as individual-orientated, early life markers signalling increased risk for schizophrenia.

Method

Previous findings using Waldrop and colleagues' MPA scale (and additional items) and systematic study of OC history are summarised concerning schizophrenia patients and individuals at heightened genetic risk for schizophrenia.

Results

Significantly increased rates of both MPA and OC are consistently found in patients with schizophrenia. Minor physical anomalies are stable characteristics over time and can be studied efficiently from early childhood onward. Minor physical anomalies predict a variety of mental disorders in normal-risk children, but the predictive efficiency of MPA for schizophrenia in genetic high-risk samples and in the general population is unknown. Obstetric complications predict serious mental disturbance and neurodisorder in genetic high-risk cases, as well as doubling or tripling the individual's risk for schizophrenia in the general population. Obstetric complication results are sensitive to methodology and are best investigated using prospectively recorded information and an efficient OC scale for scoring the information.

Conclusions

Both MPA and OC should be included in batteries of methods for identifying individuals at an increased risk for schizophrenia. However, increased rates of MPA and OC are not pathognomonic for schizophrenia, but rather characterise individuals at risk of a much broader range of mental and physical abnormality, as well as normality. Minor physical anomalies and OC are not in themselves stigmatising, but their possible identification as markers for ‘increased risk for schizophrenia’ should be used judiciously. Further research is recommended regarding the MPA and OC patterns related to schizophrenia.

Recent decades have witnessed dramatically increased scientific interest in congenital minor physical anomalies (MPA) and obstetric complications (OC) as possible individual-orientated characteristics signalling a heightened risk for schizophrenia. This interest has been promoted by major empirical findings demonstrating the importance of the prenatal, perinatal and early childhood periods; by the formulation of a corresponding neurodevelopmental theory of schizophrenia [1,2]; and by increasing consideration of possibilities for preventive intervention. Intervention at an individual level would require criteria and practicable methods for the identification of individuals with a greatly heightened risk for schizophrenia, while avoiding stigmatisation of both true positive and false positive cases. The purpose of this paper is to explore MPA and OC as practicable early risk markers for schizophrenia.

Minor physical anomalies and OC have been the focus of many studies in schizophrenia. The purpose of these studies has typically been to investigate the basic phenomenology and possible pathophysiology of this disease, as well as its developmental roots. Application of these results in the context of individual-orientated preventive intervention would raise additional important questions about: (i) sensitivity; that is, to what extent do these physical phenomena characterise individuals who will later develop schizophrenia? (ii) stability of this antecedent of schizophrenia; that is, at what ages could these characteristics be identified as antecedents, and what specific form would be evident at that time? (iii) methodological efficiency; that is, whether the phenomena can be measured efficiently in a practical context; and (iv) specificity; that is, to what extent do these phenomena characterise only individuals with (pre)schizophrenia versus other forms of (pre)psychopathology versus normality. The present study will explore these questions for both MPA and OC.

Minor physical anomalies

Minor physical anomalies are very minor structural deviations found in many areas of the body, but are typically studied in the readily visible surface areas in the head, eyes, mouth, ears, hands and feet. To our knowledge, all studies of MPA in schizophrenia have employed the MPA scale developed by Waldrop and colleagues more than 30 years ago [3]. This scale has been modified or supplemented with additional items or methods in several recent studies [4–6]. The Waldrop scale consists of 18 specific items; for example, abnormal hair whorls (head), epicanthus (eyes), high/steepled palate (mouth), adherent lobes (ears), curved fifth finger (hand), and large gap between first and second toes (feet).

Minor physical anomalies generally have little cosmetic or physiological importance in themselves [7], and typically go unnoticed to the untrained eye. An examination for MPA should be correspondingly non-stigmatising for the subject. These benign physical deviations are scientifically interesting because they represent lasting evidence of fetal maldevelopment during early pregnancy (predominantly trimester 1 and early trimester 2), which is also a time of major development of the brain. Further, the anatomic structures involved in most MPA share their embryonic substrate with the developing brain, as both arise out of cellular differentiation in the ectoderm [7]. A large number of MPA could signal parallel prenatal maldevelopment in the brain, with an increased risk for later psychopathology; that is, at least partially based on early neural maldevelopment.

Patient control differences and sensitivity to schizophrenia

Minor physical anomalies may well be the area of schizophrenia research that has shown the greatest consistency of both methodology and results across studies. To our knowledge, at least 12 different studies using the Waldrop MPA scale have found adult schizophrenia patients to show significantly increased rates of MPA compared to comparison subjects, while only one study failed to find a significant increase in patients [6]. Even in monozygotic twin pairs whose physical development is strongly influenced by within-pair genetic factors, twins with schizophrenia tended to have significantly higher MPA rates than did their mentally well cotwins [8].

Minor physical anomalies are common in the general population. Our recent study of 75 normal subjects showed 95% of them to have one or more of 41 specific MPA included in our scale, with an average of 2.73 (SD = 1.68) MPA per person [6]. In contrast, our representative sample of patients with schizophrenia had significantly increased rates of MPA both in total (mean = 6.37, SD = 2.62) and in each of the six specific body areas.

Minor physical anomalies showed reasonably high sensitivity for schizophrenia versus normality in these predefined groups. A cut-off score (> 6 MPA), which described only 5% of the normal comparison cases, characterised fully 60% of the patients with schizophrenia [6]. Logistic regression analyses, which were used to identify the specific MPA that best discriminated patients from comparison cases, found five specific items that made significant (p < 0.05) independent contributions to this discrimination. These are: (i) curved fifth finger (p = 0.001); (ii) epicanthus (p = 0.008); (iii) high/steepled palate (p = 0.01); (iv) hyperconvex fingernails (p = 0.03); and (v) thin upper lip (p = 0.05). This regression model correctly classified 73% of the patients and 85% of the comparison subjects, with an 80% correct classification in the total sample. Both total amount of MPA and a selected combination of particularly relevant MPA can thus rather efficiently discriminate patients whom are already ill from normal comparison subjects.

Workers in Dublin have undoubtedly performed the most methodologically sophisticated studies of MPA in schizophrenia [4,5,7,9]. These studies have supplemented the classic Waldrop MPA scale with anthropomorphic measurements of distances (e.g. 40 landmarks in the head and face, limbs and trunk, in order to characterise body size and shape). The resulting regression models correctly classified 85% of schizophrenia patients and 57% of normal controls in one data set [5], and 89% of male patients versus 58% male controls and 88% female patients versus 79% female controls in another data set [9]. The Dublin group has found evidence of widespread disproportionality in the body, including the lower legs, in schizophrenia. Their focus has, nevertheless, been on a particular cranio-facial profile, with ‘an overall narrowing and elongation of the mid-facial and lower facial region with widening of the skull base and extensive abnormalities of the mouth, ears and eyes’ [5; p. 1160].

Across studies, the two specific classic MPA bearing the strongest relationship to schizophrenia are found in the facial region; that is, in the mouth (high/steepled palate) and the eye (epicanthus). These specific MPA significantly characterised the patients compared to the healthy subjects in both our samples (p < 0.0001, p = 0.0003 for these two MPA, respectively) [6] and the Dublin sample (p < 0.001, p < 0.001) respectively [5].

Stability of this antecedent of schizophrenia

Approximately 2% of newborns have malformationlike ‘deformations’ (e.g. in the cranio-facial nasal, auricular and mandibular regions) that are typically caused by intrauterine pressure [10]. In most cases, these deformations are transitory in nature and resolve spontaneously within several months of birth [10].

In contrast, the classic Waldrop MPA investigated in schizophrenia reflect basic structural anomalies of the type that do not change form over time. Perhaps the major scientific advantage of MPA is that they represent externally visible ‘indelible fingerprints’ of events occurring early in gestation, which remain constant throughout life. Their use thus permits a veridical retrospective view of development during early pregnancy, but which can be studied at any time during the individual's life.

Some reservation must nevertheless be made for two Waldrop MPA; that is, epicanthus and head circumference. Epicanthus is a common and often transient developmental phenomenon in infants. It is seen in 75% of 3-months-old children [3] compared to 3% [6] to 40% [5] of normal adults. Thus, epicanthus can not be investigated as a permanent individual characteristic until early childhood [3]. Furthermore, head circumference studied in adulthood or even middle childhood may be a very unreliable measure of early prenatal cerebral development. We have found no relationship at all between head circumference at birth and in adulthood among patients with schizophrenia [11]. While five independent studies have found schizophrenia patients to have a significantly smaller head circumference at birth (typically in females) [11–15], other studies have found a disproportionately large head circumference in male schizophrenia patients in adulthood [11,16]. Beyond possible gender differences, these different results may reflect both normal and abnormal developmental phenomena. First, two-thirds of normal head growth takes place during the first two years after birth [17], and stability in relative size from prenatal to later postnatal ages may not be expected in general. Second, processes that disturb cerebral development in schizophrenia could potentially lead both to retarded development prenatally, and to hyper-development postnatally. For example, Bassett et al. have suggested that increased head size in patients could result from skull thickening or increased ventricular and sulcal-fluid due to early brain injury; from overgrowth due to pleiotropic expression of a developmental gene; or from unknown acquired (environmental) factors subsequent to the onset of schizophrenia [16].

With the exception of very early epicanthus and later head circumference, the Waldrop MPA should be able to be assessed as a stable individual characteristic at any particular subject age.

Methodological efficiency

The efficacy of MPA methodology has been a topic of continuing concern among researchers. On the positive side, the examination typically takes only 10–15 min (even with an extension to a total of 41 items which are described later). It requires the removal of only shoes and stockings, and involves minimal physical contact. These methodological characteristics are especially advantageous in the assessment of sensitive individuals.

On the other hand, the Waldrop MPA and norms are most applicable to Caucasian subjects. Little is apparently known about the appropriateness of these MPA for congenital maldevelopment among other ethnic and racial groups. Furthermore, the Waldrop scale has been criticised both on the grounds that its 18 items constitute a very limited selection among the many different MPA and congenital malformations that can be studied, and on the grounds that it requires subjective judgements and thus may be difficult and unreliable to use (e.g. [4]). The Waldrop scale items do indeed represent only a minimal selection of MPA, but these should be highly relevant for the identification of individuals with an increased risk for schizophrenia. The scale items were originally chosen on the basis of MPA found among patients with childhood schizophrenia, which was originally investigated by Goldfarb and Botstein as reported by Lane and coworkers [5].

We have empirically tested the efficiency of the Waldrop scale in schizophrenia research by comparing the 18 Waldrop items with 23 additional MPA in the same six body areas, which were chosen on the basis of paediatric sources [6]. Patients with schizophrenia showed significant increases in MPA measured by both the Waldrop scale (significant for head, eyes, ears, mouth and hand areas, and in total) and by the new items (significant for head, eyes, mouth, hand, foot, and in total). Significantly increased frequencies of MPA in patients were found somewhat more frequently on the Waldrop scale than on the new scale items, and the three specific MPA that best discriminated patients from controls came from the Waldrop scale.

Furthermore, while the Waldrop scale and the new items do require subjective judgements about the presence, and in some cases the degree, of the particular MPA, we have attained very satisfactory inter-judge agreement on MPA. Our examiners had an intraclass correlation of + 0.84 (p < 0.005) for total MPA score [6], and very satisfactory agreement (correlations > + 0.95) was also reached on quantitative measurement; for example, the degree of curved fifth finger, head circumference and interpupil distance. The MPA examiner needs no particular academic background, but must have both familiarity with the scale and ability in visual pattern recognition. In total, our experience is that MPA can be studied very reliably. In order to promote standardisation across examiners and studies, we are currently developing a detailed graphic manual to define visually the expanded scale MPA [6].

Specificity for schizophrenia

Minor physical anomalies are clearly nonspecific for schizophrenia. First, as noted earlier, one or more MPA were found in 95% of our normal comparison subjects [6]. The two specific Waldrop MPA that were found to be of greatest relevance in schizophrenia; that is, high/steepled palate and epicanthus, were also found in 53% and 40%, respectively, of normal subjects in the Dublin group's study [5], and in 19% and 3%, respectively, of normal comparison subjects who we studied [6]. The MPA that most strongly discriminated patients from normal subjects in our regression model (curved fifth finger) was also found in 41% of normal subjects [6].

Second, we found significantly increased rates of MPA in the normal siblings of patients with schizophrenia compared to comparison subjects [6]. In total, 38% of the siblings showed an increased level of MPA, which was found in only 5% of the comparison subjects. The rates of ‘key’ MPA, such as high/steepled palate and epicanthus, were identical in the patient group (52% and 24%, respectively) and the normal sibling group (52% and 23%, respectively), in spite of the total absence of intrafamilial similarity on MPA in the patients and siblings.

Third, significantly increased rates of Waldrop scale MPA have also been observed in a broad range of mental, physical and behavioural disorders in addition to schizophrenia. For example, attention deficit disorder, autism, cerebral palsy, epilepsy, fetal alcohol syndrome, hyperactivity, learning impairments, mental retardation, poor motor coordination, and speech and hearing disabilities [18,19]. In the longitudinal study we conducted, congenital malformations of a typically minor type predicted poor attention-persistence during early childhood, and increased anxiety-proneness, poor relational competency and global mental disturbance at 6 years of age among normal-risk children [20]. In contrast, malformations did not predict mental abnormality in genetic high-risk cases, for whom other risk factors had greater saliency [20]. Reduced head circumference at birth is characteristic not only of patients with schizophrenia but also of patients with schizoaffective disorder and unspecified functional psychosis [21]. Nevertheless, it remains to be seen whether a highly specific cranio-facial profile, such as that identified by the Dublin study group, or other detailed patterns among MPA would be more pathognomonic for schizophrenia.

Obstetric complications

Obstetric complications can be defined as the broad class of somatic deviations from an expected normal course of events and offspring development during the sequence from pregnancy through the early neonatal period [22]. Obstetric complications are typically divided chronologically into ‘pregnancy complications’ occurring from the time of conception to the onset of labour (e.g. pre-eclampsia, bleeding from the vaginal tract, maternal infections, etc.); ‘labour-delivery complications’ from the onset of labour until clamping of the umbilical cord after delivery (e.g. abnormal fetal presentation, prolonged or precipitous labour, instrumental delivery, etc.); and ‘neonatal complications’ from the moment of birth through the first month of postnatal life or, more typically, until discharge from the hospital (e.g. signs of hypoxia, respiratory infections, gastrointestinal disorders, etc.) [22,23].

Obstetric complications are included in the study of schizophrenia because increased rates of this type of trauma and abnormality have been observed to characterise the histories of patients with schizophrenia for at least the past four decades, and because schizophrenia is perceived as a central nervous system (CNS) disease, the basic roots of which may be substantially determined during this early developmental period. Assessment of OC will thus provide a measure of the individual's exposure to early somatic trauma and dysmorphogenesis, which is useful in understanding the development and pathophysiology of the disease or for charting the personal disease profile of the individual.

Patient-control differences, sensitivity and predictability for schizophrenia

Systematic studies conducted over at least four decades have shown rather consistently that patients with schizophrenia have significantly increased rates of OC. As summarised in 1995, significant increases in OC were found in seven of eight investigations using prospectively recorded information on OC and in nine of 13 studies using retrospective parental recall [22]. Many new and sophisticated studies have been conducted since then, using prospective OC information and patient samples ranging from approximately 75 to over 500 individuals [12,13,24–30]. Meta-analyses have also been conducted using major sample aggregations of 700 [31,32] and 854 schizophrenia cases [33], respectively. All but one [25] of these recent studies have found a significant increase in OC in the patients and/or significant relationships between OC and other characteristics of the patients.

Two major studies obtained what are generally considered to represent negative findings. In one of these studies, the Crow research group investigated OC that predicted perinatal death in the 1958 British perinatal mortality survey [34]. They found nonsignificantly increased odds ratio (OR) values of 1.4 for 35 patients with narrowly defined schizophrenia and 1.5 for 57 patients with more broadly defined schizophrenia. In the other study, Kendell and colleagues re-analysed and extended their previously ‘positive’ study [24], now finding a significant increase of generally benign OC in control cases in the series born in 1971–1974, and a significant increase of high severity OC (scored per 23), including prolonged labour and emergency caesarean section in schizophrenia patients in the series born in 1975–1978 [25]. The results of these two studies do not suggest the entire absence of a relationship between OC and schizophrenia.

Recent publications of prospective cohorts also demonstrate significant relationships between OC and the later development of schizophrenia; for example, significant relationships between offspring schizophrenia and maternal pre-eclampsia (OR = 2.4), hypertension (OR = 2.6) and ‘chronic fetal hypoxia’ (OR = 1.8) [35]. Furthermore, ‘fetal/neonatal ischemic encephalopathy’ was found to be related to later schizophrenia (relative risk (RR) = 2.1, p = 0.002) [36], while ‘perinatal brain damage’ was very strongly related to later schizophrenia (OR 7.5) [29]. The occurrence of OC also appeared to be the distinguishing differential factor for the development of schizophrenia (compared with borderline schizophrenia) within a high-risk group [37], similar to the notable OC difference observed between schizophrenia patients compared with their normal siblings [38].

Not only do OC occur more often in schizophrenia patients than the general population, but OC in these patients relate very strongly to brain characteristics of central importance to the disease. For example, recent findings indicate that labour-delivery complications (and especially prolonged labour) bear an impressively strong relationship to brain structure deviations (decreased hippocampi and increased lateral ventricles) in individuals with schizophrenia [39], with similar findings observed earlier among individuals at a high genetic risk for schizophrenia [40]. Obstetric complications were also found to be related to neuromotor abnormality in the male offspring of parents with schizophrenia in a prospective study by Marcus and colleagues [41], and to be related to the increased neurological abnormality found in the well siblings of schizophrenia patients [38,42].

The literature showing increased OC in schizophrenia patients seems all the more convincing regarding the relationship between OC and schizophrenia, as the findings represent a broad range of samples and demographic conditions; that is, both singletons and monozygotic twins; individuals adopted early in life; females and males; and offspring from all maternal parity groups, age groups and social classes; as well as patients born in numerous countries in widely spread areas in the world (North America, Europe, Scandinavia and Africa), over at least a 40-year period (1936–1979) [43]. The labour–delivery complications occurring in patients and genetic risk cases do not appear to be secondary consequences of pre-existing fetal abnormality [44], and these OC appear to make an aetiological contribution in their own right. Hence, our conclusion is that the relationship between OC and subsequent schizophrenia represents an enduring biological phenomenon of aetiological importance that is independent of the specific time period, geographical location, social context or medical practice.

In contrast, less clarity exists concerning the actual strength of the OC–schizophrenia relationship and the sensitivity with which OC identify future cases of schizophrenia. Obstetric complications may well show greater saliency in some samples than others; for example, in twin rather than singleton patients, in patients with lower age of disease onset [33] and in patients with a negative family history for schizophrenia [22]. The exact frequency with which patients have experienced OC has varied greatly across studies. For example, in studies performed before 1995, one or more OC were found in 19–87% of the patients and 14–80% of the comparison cases across studies with prospective information, and in 17–91% of the patients and 5–57% of the controls across the retrospective studies [22]. Notable variations in the absolute frequency of specific OC were also found across 12 previous studies employing the very same OC scale [32], and similar variations occurred across the newer studies published in the past five years.

Obstetric complication rates per se may vary as a function of the general frequency of OC in the base population; the source and completeness of the OC information; the operational definition of conditions to be scored as OC; and the actual manner of scoring these [43]. Hence, the OR or RR for OC in patients versus comparison cases in the same study are of considerable interest.

In a major meta-analysis of 700 patients, Geddes and Lawrie found a pooled OR of 2.0 (95% confidence interval = 1.6–2.4) for OC exposure scored as being present compared to absent across different OC [31]. Further, the mean OR or RR was 3.1 for the OC (e.g. maternal pre-eclampsia, hypertension, bleeding, diabetes, offspring hypoxia, instrumental delivery, deviant birthweight and/or gestational age, perinatal brain damage, etc.), which showed significant increases in patients in recently published studies [43]. During one investigation, we found an OR of 1.3 for a continuous score for the total number of OC, thus showing the schizophrenia risk to increase by 30% for each additional OC [45].

As a whole, these results would suggest that the risk of schizophrenia is approximately doubled or tripled if the individual is exposed to at least one of a range of different OC, with the risk increasing for each individual OC scored according to the scale we proposed [23]. According to Geddes and Lawrie's calculation, approximately one-fifth (22%) of schizophrenia cases are attributable to the occurrence of one or more of the Lewis–Murray scale OC (assuming these are aetiological) [31]; while Jones and colleagues' findings in Finland indicate that about 7% of schizophrenia in the population is attributable to their variable ‘perinatal brain damage’ [29]. These estimates are interesting but must be considered to be tentative, given the variations in (relative) rates across different studies.

No single OC has been found to be of central importance for schizophrenia, and in the words of Hultman and coworkers, ‘many different, unspecific OC, rather than any particular OC, appear to contribute to the increased rate among schizophrenics’ [12, p. 132]. Our findings also indicate that it is the total number of relevant OC, rather than the mere presence of OC, that differentiates schizophrenia patients from comparison subjects. A scoring approach that reflects the total number of OC with potential damage to the offspring thus seems to be optimal.

‘Stability’ and methodological efficiency for this antecedent of schizophrenia

By definition, an individual's true OC history is a ‘stable’ phenomenon, but the operational measure of it may be highly influenced by the source of information and the choice of system for scoring this information. Prospectively recorded OC information available in medical records or prospective cohort studies is the source of choice, but researchers and clinicians are often relegated to using retrospective parental recall of OC. While one empirical study found agreement between medical records and retrospective recall [46], we have found that retrospective recall is associated with omissive errors and tends to miss patient–control OC differences [47].

Similarly, several different systems exist for scoring OC information, and studies we have carried out show that the particular OC scale used in a study influences both the extent to which patients differ from comparison subjects [48] and the relationships that OC bear to other presumed aetiological factors, such as family history of psychosis and season of birth [49].

In a direct empirical test, the scale that we have proposed [23] proved to be most sensitive to OC differences between patients and comparison subjects [48]. This scale consists of several hundred specific OC, which cover pregnancy through to the neonatal period, and each OC is graded according to a 6-point scale reflecting the inferred probability of damage to the offspring. Experience shows that untrained scorers can learn to use this scale efficiently within two days' practise, either on their own or preferably with training assistance. The scorer needs to have acquired some basic knowledge of obstetrics and its terminology, but need not necessarily have formal medical training. Parallel and independent scoring of the same obstetric records or information by two researchers is advantageous, at least initially, to standardise procedures, test and maximise scoring reliability and avoid missing relevant information. Recording and scoring of OC should definitely be done blindly with respect to the other characteristic(s) that will be investigated in relation to OC.

For ethical reasons, permission is typically obtained either from the offspring or the mothers for accessing and using the OC information. Our experience is that these subjects are almost always interested in their own OC history and its possible relevance for illness, and that the information regarding OC is not stigmatising for the subjects or guilt-producing for the mothers.

Specificity for schizophrenia

As was the case for MPA, OC are clearly nonspecific for schizophrenia. Rates of OC vastly exceed the rate of schizophrenia 20 years later in any given general population. The well-known studies of OC by Pasamanick and Knobloch more than 40 years ago, led to the concept of ‘a continuum of reproductive casualty’ that is associated with many different types of mental and physical abnormalities, such as behaviour disorders, reading disabilities, mental deficiency, epilepsy, cerebral palsy and tics [50]. Obstetric complications continue to show a relationship to childhood and adolescent behaviour disorders [51]. The recent, prolific work of Barker and associates [52] has done much to extend the focus of OC as antecedents of a range of adult medical diseases, such as cardiovascular disease, diabetes, renal diseases and ovarian cancer.

Conclusion

Increased rates of MPA and OC clearly represent early life antecedents of both schizophrenia and many other mental disorders, as well as normality. Increased rates of MPA consistently characterise schizophrenia groups, but the efficacy of MPA as a predictor of schizophrenia within genetic high-risk groups and within unselected samples in the population is unknown. Minor physical anomalies are stable over time and can be studied efficiently from childhood onward, preferably using an extended method or a method tapping a schizophrenia-related cranio-facial profile. Obstetric complications consistently characterise schizophrenia groups and also predict serious subsequent mental disturbance, structural brain deficits and neurobehavioural abnormality among children with genetic risk for schizophrenia. In the general population, the occurrence of one or more OC may double or triple the individual's risk for schizophrenia, with increasing risk increments as OC multiply. Approximately one-fifth of schizophrenia cases may be due to OC. Obstetric complications results are sensitive to methodology and are best studied using prospectively recorded information and an efficient scoring scale. The study of MPA and OC is not in itself stigmatising. Nevertheless, the identification of these markers as indicating ‘risk for schizophrenia’ or other mental disturbance could potentially be emotionally problematic for the individuals in question, and such identification must be used judiciously. Further research needs to be done on the characteristic patterns of both MPA and OC that are related to schizophrenia, as well as the possible means of structuring environments to ameliorate the subsequent mental and physical development of individual children who possess these markers of increased risk for serious psychopathology.