- Split View
-
Views
-
Cite
Cite
Unn K. Haukvik, Lars M. Rimol, J. Cooper Roddey, Cecilie B. Hartberg, Elisabeth H. Lange, Anja Vaskinn, Ingrid Melle, Ole A. Andreassen, Anders Dale, Ingrid Agartz, Normal Birth Weight Variation Is Related to Cortical Morphology Across the Psychosis Spectrum, Schizophrenia Bulletin, Volume 40, Issue 2, March 2014, Pages 410–419, https://doi.org/10.1093/schbul/sbt005
- Share Icon Share
Abstract
Normal birth weight variation affects schizophrenia risk and cognitive performance in schizophrenia patients and healthy controls. Brain cortical anatomy is altered in psychotic disorders and in low birth weight subjects, but if birth weight variation relates to cortical morphology across the psychosis spectrum is not known.
Magnetic Resonance Imaging brain scans and clinical-, neurocognitive-, and medical birth registry data were collected from 359 adults including patients with a DSM-IV diagnosis of schizophrenia (n = 90, mean age 29.4±10.2 [95% CI], 62% male), bipolar disorder (n = 79, age 29.4±11.8, 39% male) or other psychosis (n = 40, age 26.3±10.0, 56% male), and healthy controls (n = 140, age 30.8±12.0,53% male). We explored the relationship between whole-range birth weight variation and cortical surface area and thickness and their possible associations to cognitive performance.
Across all groups, lower birth weight was associated with smaller total surface area (t = 3.87, P = .0001), within specific regions of the temporal, parietal, and frontal cortex bilaterally. There were no associations between birth weight and cortical thickness, and no diagnosis by birth weight interaction effects on cortical thickness or surface area. Smaller cortical area (t = 2.50, P = .013) and lower birth weight (t = 2.53, P = .012) were significantly related to poorer working memory performance in all diagnostic groups except schizophrenia.
Birth weight relates to adult cortical surface area, but not cortical thickness, in patients across the psychosis spectrum and in healthy controls. Cortical area appears to be a diagnosis-independent general marker of early neurodevelopment, with a dose-response association to normal birth weight variation.
Introduction
Epidemiologic studies suggest a fetal origin of mental illnesses such as autism, personality disorders, and schizophrenia.1 According to the neurodevelopmental hypothesis of schizophrenia, obstetric complications (OCs) may cause subtle alterations in neurodevelopment that leave the brain susceptible to developing schizophrenia.2 Low birth weight (LBW; <2500g)3 reflects an array of different OCs, eg, hypoxia, prematurity, and intrauterine growth retardation (IUGR), all of which may increase schizophrenia risk.4,5 LBW has also been associated with other psychiatric disorders, including bipolar disorder.6 Recent epidemiological findings, however, suggest that the increased risk for mental disorders is not restricted to LBW subjects. Instead, the OR for schizophrenia appears to decrease linearly across the whole birth weight (BW) range (500–4500g), from OR 1.73 for the smallest newborns (BW below 1499g) to OR 0.97 for the largest (BW above 4500g).7 A similar pattern is observed in affective disorders, with ORs decreasing from 1.84 for the smallest newborns to 0.76 for the largest.7 BW may represent an indirect, continuous measure of severe and subtle prenatal neurodevelopmental adversities of importance to the risk for schizophrenia and other psychotic disorders.
Magnetic Resonance Imaging (MRI) studies of subjects with LBW and very low birth weight (VLBW, <1500g) have demonstrated reduced cortical gray matter volume8 and cortical surface area and thickness9 in otherwise healthy adolescents. These effects may not be limited to subjects with VLBW or LBW. Intrauterine complications of varying severity (eg, twinning and IUGR) have been demonstrated to affect prenatal neurodevelopment,10 and animal models have shown that a wide array of subtle to more severe prenatal complications, including infection and chronic hypoxia, may cause abnormalities in cortical morphology.11 It is biologically plausible that subtle and severe prenatal adversities relate to cortical morphology in a continuous (dose-response), rather than dichotomous (cut-off), fashion. This was recently demonstrated in a twin study of healthy children and young adults12 but remains unexplored in psychotic disorders.
Abnormal prenatal brain development may have distinct effects on cortical surface area and thickness. According to the radial unit hypothesis of cortical neurogenesis, neural stem cells in the ventricular zone divide symmetrically before cortical migration, and changes in the proliferation rate may increase the number of radial columnar units and cause an increase in cortical surface area but not thickness.13,14 Similar changes affecting the asymmetrical division of the neuronal founder cells lead to an increased number of cells within a radial column but do not affect cortical surface area.13 Moreover, there is a substantial growth in cortical area during the third trimester. Interestingly, schizophrenia, and to some extent bipolar disorder, have been associated with regionally specific and distinct reductions in cortical thickness and area,15,16 but the developmental context of these changes in cortical morphology has not been established. Because the risk for psychotic disorders has been reported to vary in a dose-response relationship with normal BW variation,7 normal BW variation could be related to the cortical abnormalities observed in psychotic disorders.
Of importance to everyday functioning, BW may influence cognitive performance. In healthy adults, BW across the normal range (2500–4500g) has been related to IQ measurements,17,18 and in patients with schizophrenia, normal BW variation has been related to executive functioning and working memory performance.20 Impaired working memory is an established trait in psychotic disorders,20 but has also been associated with altered cortical gray matter integrity independent of a diagnosis of psychosis.21 Working memory performance involves broad brain networks within the frontoparietal circuitry and in particular the dorsolateral prefrontal cortex.21 Given these links among normal BW variation, cognitive impairments, and abnormal brain morphology, it is possible that altered cortical morphology mediates BW effects on cognitive functioning.
The aims of this study were (1) to examine associations between whole-range BW variation and cortical surface area and thickness, in healthy subjects and patients with diagnoses across the psychosis spectrum and (2) to explore whether such associations, if any, were related to cognitive functioning.
We hypothesized that BW would associate with cortical area and thickness independent of a diagnosis of psychosis. We also hypothesized that if cortical thickness or area were associated with BW, they would act as mediators of the effect of BW on cognitive functioning, with diagnosis-specific effects, because cognitive impairments are trait markers in psychotic disorders.20
This is, to our knowledge, the first study to investigate the effects of normal BW variation on distinct cortical surface area and thickness measures in patients across the psychosis spectrum and the possible mediating effects on cognitive functioning.
Methods
Subjects
The subject sample (n = 359) consisted of patients with a DSM-IV diagnosis within the schizophrenia spectrum, schizophrenia (DSM-IV 295.1, 295.3, 295.6, and 295.9) (n = 75), schizophreniform disorder (DSM-IV 295.4) (n = 8), or schizoaffective disorder (DSM-IV 295.7) (n = 7); the bipolar spectrum, Bipolar I disorder (DSM-IV 296.0–7) (n = 47), Bipolar II disorder (DSM-IV 296.89) (n = 28), or bipolar disorder not otherwise specified (NOS; DSM-IV 296.80) (n = 4); and other psychoses, depressive psychosis (DSM-IV 286.2-3) (n = 13) or psychosis NOS (DSM-IV 298.9) (n = 37), and healthy controls (n = 140) from the on-going multicenter Thematically Organized Psychosis (TOP) Study at the University of Oslo, Norway. Of the subjects in this sample, 280 were a part of previous reports of cortical area and thickness.15,22 All subjects were born in Norway and were between 18 and 42 years. Five subjects were of non-Caucasian origin, and 14 subjects were of mixed origin (with 1 Caucasian parent).
Patients were included from 4 major psychiatric hospitals and their outpatient clinics that together cover most of the population in Oslo. The patient inclusion criteria were present or previous psychotic episode and lack of any current or previous somatic illness affecting brain morphology. All patients underwent thorough clinical investigation by specially trained psychologists and physicians. Clinical diagnoses were assessed using the Structured Clinical Interview for DSM-IV axis 1 disorder (SCID-I) module A-E23, with an overall agreement for diagnostic categories of 82%, kappa = 0.77 (95% CI 0.60–0.94). Psychosocial function was assessed with the Global Assessment of Function (GAF) scale, split version. Current psychotic symptoms were rated by the use of the Positive and Negative Syndrome Scale (PANSS), with intraclass coefficients of 0.73 and 0.86 for the positive and negative subscales, respectively.24 Of the patients, 90 were first-episode psychotic patients (defined as having received medical treatment for any psychotic condition for less than a year).
The healthy control subjects were randomly selected from the national population register and resident in the same catchment area as the patients. They were interviewed by trained psychologists and examined with the Primary Care Evaluation of Mental Disorders (Prime-MD)25 to ensure no current or previous psychiatric disorders. Control subjects with current or previous somatic illness that could affect brain morphology or substance misuse disorders were excluded. Demographic and clinical variables are listed in table 1.
. | Control Subjects, n = 140 . | Schizophrenia Spectrum, n = 90 . | Bipolar Spectrum, n = 79 . | Other Psychoses, n = 50 . | Statistics, P-Values . | ||||
---|---|---|---|---|---|---|---|---|---|
Number . | % . | Number . | % . | Number . | % . | Number . | % . | Chi square . | |
Sex (m/f) | 74/66 | 53/47 | 56/34 | 62/38 | 31/48 | 39/61 | 28/22 | 56/44 | .023 |
Handedness (r/l/a) n = 351 | 130/9/1 | 93/6/1 | 75/10/1 | 87/12/1 | 67/8/1 | 88/11/1 | 43/6/0 | 88/12/0 | ns |
Mean (SD) | Range | Mean (SD) | Range | Mean (SD) | Range | Mean (SD) | Range | ANOVA | |
Age (y) | 30.8 (6.1) | 18–42 | 29.4 (5.2) | 20–41 | 29.4 (6.0) | 19–42 | 26.3 (5.1) | 19–40 | <.0005 |
Gestational age (wk) | 39.9 (1.8) | 35–45 | 39.7 (1.8) | 33–44 | 39.9 (2.2) | 31–44 | 38.9 (3.2) | 24–44 | .042 |
Maternal age (y) | 27.3 (4.8) | 18–40 | 27.8 (5.4) | 17–41 | 27.9 (5.2) | 18–40 | 28.0 (6.1) | 17–41 | ns |
Birth weight (g) | 3518 (528) | 2200–4990 | 3440 (540) | 1940–4560 | 3552 (560) | 1880–4900 | 3416 (656) | 600–4600 | ns |
Birth head circumference (cm), n = 176 | 35.3 (1.4) | 32–38 | 35.4 (2.8) | 32–38 | 35.1 (1.4) | 32–38 | 35.4 (1.3) | 33–38 | ns |
Years of education | 14.1 (2.2) | 9–20 | 12.7 (2.3) | 9–18 | 13.6 (2.3) | 9–20 | 12.3 (2.3) | 7–18 | <.0005 |
Premorbid IQ, n = 337 | 107.0 (3.7) | 97.7–114.7 | 105.2 (4.5) | 91.7–113.7 | 106.3 (3.8) | 97.6–114.7 | 105.0 (3.8) | 93.8–111.8 | .001 |
Working memory, n = 350 | 16.5 (3.6) | 9–26 | 14.5 (3.2) | 8–23 | 16.1 (3.5) | 10–25 | 15.5 (3.9) | 10–26 | .001 |
GAF symptom | — | — | 42.4 (11.4) | — | 57.0 (9.8) | — | 46.4 (13.9) | — | <.0005 |
GAF function | — | — | 43.0 (10.4) | — | 55.6 (12.7) | — | 48.9 (14.5) | — | <.0005 |
PANSS total | — | — | 61.9 (16.2) | — | 46.1 (10.3) | — | 55.4 (12.6) | — | <.0005 |
Age at illness onset | — | — | 24.4 (4.9) | — | 23.8 (6.0) | — | 23.6 (5.0) | — | ns |
Medication (DDD), n = 194 | |||||||||
Antipsychotics, n = 153 | — | — | 1.5 (1.2) | — | 0.9 (0.7) | — | 1.1 (0.9) | — | na |
Antiepileptics, n = 64 | — | — | 0.7 (0.4) | — | 0.7 (0.4) | — | 0.5 (0.4) | — | na |
Lithium, n = 12 | — | — | — | — | 1.0 (0.3) | — | — | — | na |
Antidepressant, n = 81 | — | — | 1.7 (1.1) | — | 1.6 (0.8) | — | 1.6 (1.4) | — | na |
. | Control Subjects, n = 140 . | Schizophrenia Spectrum, n = 90 . | Bipolar Spectrum, n = 79 . | Other Psychoses, n = 50 . | Statistics, P-Values . | ||||
---|---|---|---|---|---|---|---|---|---|
Number . | % . | Number . | % . | Number . | % . | Number . | % . | Chi square . | |
Sex (m/f) | 74/66 | 53/47 | 56/34 | 62/38 | 31/48 | 39/61 | 28/22 | 56/44 | .023 |
Handedness (r/l/a) n = 351 | 130/9/1 | 93/6/1 | 75/10/1 | 87/12/1 | 67/8/1 | 88/11/1 | 43/6/0 | 88/12/0 | ns |
Mean (SD) | Range | Mean (SD) | Range | Mean (SD) | Range | Mean (SD) | Range | ANOVA | |
Age (y) | 30.8 (6.1) | 18–42 | 29.4 (5.2) | 20–41 | 29.4 (6.0) | 19–42 | 26.3 (5.1) | 19–40 | <.0005 |
Gestational age (wk) | 39.9 (1.8) | 35–45 | 39.7 (1.8) | 33–44 | 39.9 (2.2) | 31–44 | 38.9 (3.2) | 24–44 | .042 |
Maternal age (y) | 27.3 (4.8) | 18–40 | 27.8 (5.4) | 17–41 | 27.9 (5.2) | 18–40 | 28.0 (6.1) | 17–41 | ns |
Birth weight (g) | 3518 (528) | 2200–4990 | 3440 (540) | 1940–4560 | 3552 (560) | 1880–4900 | 3416 (656) | 600–4600 | ns |
Birth head circumference (cm), n = 176 | 35.3 (1.4) | 32–38 | 35.4 (2.8) | 32–38 | 35.1 (1.4) | 32–38 | 35.4 (1.3) | 33–38 | ns |
Years of education | 14.1 (2.2) | 9–20 | 12.7 (2.3) | 9–18 | 13.6 (2.3) | 9–20 | 12.3 (2.3) | 7–18 | <.0005 |
Premorbid IQ, n = 337 | 107.0 (3.7) | 97.7–114.7 | 105.2 (4.5) | 91.7–113.7 | 106.3 (3.8) | 97.6–114.7 | 105.0 (3.8) | 93.8–111.8 | .001 |
Working memory, n = 350 | 16.5 (3.6) | 9–26 | 14.5 (3.2) | 8–23 | 16.1 (3.5) | 10–25 | 15.5 (3.9) | 10–26 | .001 |
GAF symptom | — | — | 42.4 (11.4) | — | 57.0 (9.8) | — | 46.4 (13.9) | — | <.0005 |
GAF function | — | — | 43.0 (10.4) | — | 55.6 (12.7) | — | 48.9 (14.5) | — | <.0005 |
PANSS total | — | — | 61.9 (16.2) | — | 46.1 (10.3) | — | 55.4 (12.6) | — | <.0005 |
Age at illness onset | — | — | 24.4 (4.9) | — | 23.8 (6.0) | — | 23.6 (5.0) | — | ns |
Medication (DDD), n = 194 | |||||||||
Antipsychotics, n = 153 | — | — | 1.5 (1.2) | — | 0.9 (0.7) | — | 1.1 (0.9) | — | na |
Antiepileptics, n = 64 | — | — | 0.7 (0.4) | — | 0.7 (0.4) | — | 0.5 (0.4) | — | na |
Lithium, n = 12 | — | — | — | — | 1.0 (0.3) | — | — | — | na |
Antidepressant, n = 81 | — | — | 1.7 (1.1) | — | 1.6 (0.8) | — | 1.6 (1.4) | — | na |
Note: Premorbid IQ, National Adult Reading Test IQ estimate; Working memory, digit span sum of forward and backward trials; GAF, Global assessment of function; PANSS, Positive and negative syndrome scale; DDD, defined daily dosage (World Health Organization Collaborating Center for Drug Statistics Methodology [hhtp://www.whocc.no/atcdd]); ns, not significant; na, not applicable.
. | Control Subjects, n = 140 . | Schizophrenia Spectrum, n = 90 . | Bipolar Spectrum, n = 79 . | Other Psychoses, n = 50 . | Statistics, P-Values . | ||||
---|---|---|---|---|---|---|---|---|---|
Number . | % . | Number . | % . | Number . | % . | Number . | % . | Chi square . | |
Sex (m/f) | 74/66 | 53/47 | 56/34 | 62/38 | 31/48 | 39/61 | 28/22 | 56/44 | .023 |
Handedness (r/l/a) n = 351 | 130/9/1 | 93/6/1 | 75/10/1 | 87/12/1 | 67/8/1 | 88/11/1 | 43/6/0 | 88/12/0 | ns |
Mean (SD) | Range | Mean (SD) | Range | Mean (SD) | Range | Mean (SD) | Range | ANOVA | |
Age (y) | 30.8 (6.1) | 18–42 | 29.4 (5.2) | 20–41 | 29.4 (6.0) | 19–42 | 26.3 (5.1) | 19–40 | <.0005 |
Gestational age (wk) | 39.9 (1.8) | 35–45 | 39.7 (1.8) | 33–44 | 39.9 (2.2) | 31–44 | 38.9 (3.2) | 24–44 | .042 |
Maternal age (y) | 27.3 (4.8) | 18–40 | 27.8 (5.4) | 17–41 | 27.9 (5.2) | 18–40 | 28.0 (6.1) | 17–41 | ns |
Birth weight (g) | 3518 (528) | 2200–4990 | 3440 (540) | 1940–4560 | 3552 (560) | 1880–4900 | 3416 (656) | 600–4600 | ns |
Birth head circumference (cm), n = 176 | 35.3 (1.4) | 32–38 | 35.4 (2.8) | 32–38 | 35.1 (1.4) | 32–38 | 35.4 (1.3) | 33–38 | ns |
Years of education | 14.1 (2.2) | 9–20 | 12.7 (2.3) | 9–18 | 13.6 (2.3) | 9–20 | 12.3 (2.3) | 7–18 | <.0005 |
Premorbid IQ, n = 337 | 107.0 (3.7) | 97.7–114.7 | 105.2 (4.5) | 91.7–113.7 | 106.3 (3.8) | 97.6–114.7 | 105.0 (3.8) | 93.8–111.8 | .001 |
Working memory, n = 350 | 16.5 (3.6) | 9–26 | 14.5 (3.2) | 8–23 | 16.1 (3.5) | 10–25 | 15.5 (3.9) | 10–26 | .001 |
GAF symptom | — | — | 42.4 (11.4) | — | 57.0 (9.8) | — | 46.4 (13.9) | — | <.0005 |
GAF function | — | — | 43.0 (10.4) | — | 55.6 (12.7) | — | 48.9 (14.5) | — | <.0005 |
PANSS total | — | — | 61.9 (16.2) | — | 46.1 (10.3) | — | 55.4 (12.6) | — | <.0005 |
Age at illness onset | — | — | 24.4 (4.9) | — | 23.8 (6.0) | — | 23.6 (5.0) | — | ns |
Medication (DDD), n = 194 | |||||||||
Antipsychotics, n = 153 | — | — | 1.5 (1.2) | — | 0.9 (0.7) | — | 1.1 (0.9) | — | na |
Antiepileptics, n = 64 | — | — | 0.7 (0.4) | — | 0.7 (0.4) | — | 0.5 (0.4) | — | na |
Lithium, n = 12 | — | — | — | — | 1.0 (0.3) | — | — | — | na |
Antidepressant, n = 81 | — | — | 1.7 (1.1) | — | 1.6 (0.8) | — | 1.6 (1.4) | — | na |
. | Control Subjects, n = 140 . | Schizophrenia Spectrum, n = 90 . | Bipolar Spectrum, n = 79 . | Other Psychoses, n = 50 . | Statistics, P-Values . | ||||
---|---|---|---|---|---|---|---|---|---|
Number . | % . | Number . | % . | Number . | % . | Number . | % . | Chi square . | |
Sex (m/f) | 74/66 | 53/47 | 56/34 | 62/38 | 31/48 | 39/61 | 28/22 | 56/44 | .023 |
Handedness (r/l/a) n = 351 | 130/9/1 | 93/6/1 | 75/10/1 | 87/12/1 | 67/8/1 | 88/11/1 | 43/6/0 | 88/12/0 | ns |
Mean (SD) | Range | Mean (SD) | Range | Mean (SD) | Range | Mean (SD) | Range | ANOVA | |
Age (y) | 30.8 (6.1) | 18–42 | 29.4 (5.2) | 20–41 | 29.4 (6.0) | 19–42 | 26.3 (5.1) | 19–40 | <.0005 |
Gestational age (wk) | 39.9 (1.8) | 35–45 | 39.7 (1.8) | 33–44 | 39.9 (2.2) | 31–44 | 38.9 (3.2) | 24–44 | .042 |
Maternal age (y) | 27.3 (4.8) | 18–40 | 27.8 (5.4) | 17–41 | 27.9 (5.2) | 18–40 | 28.0 (6.1) | 17–41 | ns |
Birth weight (g) | 3518 (528) | 2200–4990 | 3440 (540) | 1940–4560 | 3552 (560) | 1880–4900 | 3416 (656) | 600–4600 | ns |
Birth head circumference (cm), n = 176 | 35.3 (1.4) | 32–38 | 35.4 (2.8) | 32–38 | 35.1 (1.4) | 32–38 | 35.4 (1.3) | 33–38 | ns |
Years of education | 14.1 (2.2) | 9–20 | 12.7 (2.3) | 9–18 | 13.6 (2.3) | 9–20 | 12.3 (2.3) | 7–18 | <.0005 |
Premorbid IQ, n = 337 | 107.0 (3.7) | 97.7–114.7 | 105.2 (4.5) | 91.7–113.7 | 106.3 (3.8) | 97.6–114.7 | 105.0 (3.8) | 93.8–111.8 | .001 |
Working memory, n = 350 | 16.5 (3.6) | 9–26 | 14.5 (3.2) | 8–23 | 16.1 (3.5) | 10–25 | 15.5 (3.9) | 10–26 | .001 |
GAF symptom | — | — | 42.4 (11.4) | — | 57.0 (9.8) | — | 46.4 (13.9) | — | <.0005 |
GAF function | — | — | 43.0 (10.4) | — | 55.6 (12.7) | — | 48.9 (14.5) | — | <.0005 |
PANSS total | — | — | 61.9 (16.2) | — | 46.1 (10.3) | — | 55.4 (12.6) | — | <.0005 |
Age at illness onset | — | — | 24.4 (4.9) | — | 23.8 (6.0) | — | 23.6 (5.0) | — | ns |
Medication (DDD), n = 194 | |||||||||
Antipsychotics, n = 153 | — | — | 1.5 (1.2) | — | 0.9 (0.7) | — | 1.1 (0.9) | — | na |
Antiepileptics, n = 64 | — | — | 0.7 (0.4) | — | 0.7 (0.4) | — | 0.5 (0.4) | — | na |
Lithium, n = 12 | — | — | — | — | 1.0 (0.3) | — | — | — | na |
Antidepressant, n = 81 | — | — | 1.7 (1.1) | — | 1.6 (0.8) | — | 1.6 (1.4) | — | na |
Note: Premorbid IQ, National Adult Reading Test IQ estimate; Working memory, digit span sum of forward and backward trials; GAF, Global assessment of function; PANSS, Positive and negative syndrome scale; DDD, defined daily dosage (World Health Organization Collaborating Center for Drug Statistics Methodology [hhtp://www.whocc.no/atcdd]); ns, not significant; na, not applicable.
The study was approved by the Regional Committee for Medical Research Ethics and the Norwegian Data Inspectorate and was conducted in accordance with the Helsinki declaration. After complete description of the study to the subjects, written informed consent was obtained from all participating subjects.
Birth Data
The birth data were collected from the Medical Birth Registry of Norway (MBRN) that dates back to 1967. In Norway, all births after gestational week 16 are compulsorily reported to MBRN by the attending midwife or physician.26 Medical and sociodemographic data pertaining to the pregnancy, birth, and the health of the mother is recorded on a standardized notification form. This notification form has remained unchanged during the period in which all subjects in this study were born (1967–1991). Moreover, all mothers bring a standard antenatal medical record to the birth clinic, and the attending midwife or physician transfers this information to the MBRN notification form. Birth-related variables are listed in table 1.
MRI Acquisition
All participants underwent MRI scanning on a 1.5 T Siemens Magnetom Sonata scanner (Siemens Medical Solutions) equipped with a standard head coil. Two sagittal T1-weighted magnetization-prepared rapid gradient echo volumes were acquired with the Siemens tfl3d1_ns pulse sequence (TE = 3.93ms, TR = 2730ms, TI = 1000ms, flip angle = 7°; FOV = 24cm, voxel size = 1.33×0.94×1mm3, and number of partitions = 160) and subsequently averaged together, after rigid-body registration, to increase the signal to noise ratio. There was no scanner upgrade during the study period, and patients and controls were scanned consecutively.
MRI Postprocessing
The Freesurfer software (version 4.5.0; http://surfer.nmr.mgh.harvard.edu) was used to estimate cortical area and thickness. A 3-D model of the cortical surface was created by using image intensities and continuity information from the entire MR volume to construct representations of the gray/white matter boundary and pial surface.27,28 The surfaces were averaged across participants using a nonrigid high-dimensional spherical averaging method to align cortical folding patterns. Each hemisphere surface consisted of approximately 160000 vertices arranged in a triangular grid. Estimates of cortical area were obtained by computing the area of each triangle in a standardized, spherical atlas space surface tessellation. Vertexwise estimates of relative areal expansion for each subject in atlas space were then computed by assigning one-third of the area of each triangle to each of its vertices. Cortical thickness was measured as the distance between the gray/white matter boundary and the pial surface at each vertex.29 The maps were not restricted to the voxel resolution of the original data and could detect submillimeter differences between groups. The procedures were fully automated. All scans were visually inspected following standard procedures. Cortical surface area and average thickness for 68 predefined regions-of-interest (ROIs) covering the whole brain30 were extracted for statistical analyses.
Neurocognitive Assessment
All subjects with Norwegian or another Scandinavian language as their first language underwent neuropsychological assessment (n = 350). Trained psychologists used a comprehensive test battery administered in a fixed order (see Simonsen et al., 201131 for details). Working memory performance was estimated by a combined measure based on the raw scores from the Digit Span (forward trials and backward trials) and the Letter Number Span subtests from the Wechsler Adult Intelligence Scale III (WAIS-III).32 Higher raw scores correspond to better working memory performance. Raw scores were transformed to standardized z-scores normalized after the control group (mean and SD), and averaged to 1 working memory domain variable for the statistical analyses. Based on previous literature for comparison purposes,18 and because psychosis may affect current IQ estimates, IQ was estimated by the Norwegian version of the National Adult Reading Test (NART).33 NART was only administered to participants with Norwegian as their mother tongue. The NART premorbid IQ estimate was then calculated based on NART total number of errors, age, and years of education.
Statistical Analyses
Statistical analyses were performed within the statistical package SPSS version 19 (SPSS Inc.), and MATLAB version 7.9.0 (the MathWorks).
Group differences in demographic, clinical, obstetric, and neurocognitive variables were analyzed with ANOVA and chi-square statistics. All statistical analyses were two-tailed.
For the ROI analyses, a hierarchical linear regression model was used, with surface area or mean thickness as the dependent variable and age, sex, and diagnosis as independent variables entered in the first block and BW in the second block. To adjust for tests of multiple ROIs, Bonferroni correction was applied; the corrected threshold was set to P = .00076 (alpha level 0.05/68 tests [number of ROIs]). The ROIs that were significantly related to BW in the main analyses were also analyzed for between-group differences using an ANCOVA model.
Cortical thickness and area maps showing the effects of BW with age, sex, and diagnosis as covariates at over 160 000 vertices in each hemisphere were generated to explore associations not restricted to the ROIs. Vertexwise cortical surface and thickness p-maps contrasting each patient group (schizophrenia, bipolar, or other psychosis groups) with healthy controls were created with linear regression models, corrected for age and sex. For these maps, a 5% false discovery rate (FDR)34 was applied to adjust for multiple comparisons.
Linear regression analyses were conducted to search for effects of total cortical area and BW (see results below) on working memory performance and IQ. Residual analyses showed 1 influential case, this case was removed from the analyses. A simple mediation model was tested,35 with BW as the independent variable, total cortical area as the mediator, and neurocognitive test as the outcome. The raw unstandardized regression coefficients and their standard errors were tested with the Sobel method for significant mediation effects. Scatter plots revealed an opposite direction between BW and working memory in schizophrenia patents compared with the other diagnostic groups (figure 3), which suggests the presence of moderated mediation effects.35 Patients with schizophrenia also showed an opposite direction of the association between cortical area and IQ than the other diagnostic groups. Accordingly, the neurocognitive analyses were stratified by diagnosis with schizophrenia patients being analyzed separately and the other diagnostic groups combined.
Confounding Variables.
Regression analyses were conducted to explore possible confounding birth-related variables, ie, gestational age, maternal age, paternal age, parity, perinatal asphyxia, and caesarean section. Because antipsychotic medication may affect cortical morphology, we also studied the effects of defined daily dosages (calculated in accordance with the guidelines from the World Health Organization Collaborating Center for Drug Statistics Methodology [http://www.whocc.no/atcddd]) of current antipsychotic medication, generation of current antipsychotic medication, cumulative dosage of antipsychotic mediation use over time, duration of illness, and age at illness onset. When included in the regression model, none of the potential confounders had an effect on cortical thickness or area or the association between BW and cortical parameters. Moreover, we did not find ethnicity to affect the results. Accordingly, these variables were left out of the main statistical analyses.
Results
Demographic, Clinical, Neurocognitive, and Obstetric Variables
The groups differed in age, gestational age, sex distribution, education, PANSS, GAF, premorbid IQ, and working memory (table 1).
Cortical Surface Area and Thickness
Lower BW was related to smaller surface area in the left middle temporal cortex, the left rostral anterior cingulate, the right inferior parietal cortex, the right parahippocampal cortex, and the right rostral anterior cingulate (table 2, see online supplementary figure 1), as well as total surface area (t(352) = 3.87, P = .0001; figure 1), when age, sex, and diagnosis were corrected for. The FDR-corrected cortical maps showed a more widespread relationship between BW and cortical area, in particular the middle temporal lobe, the anterior cingulate, frontal cortex, and parietal cortex (figure 2). Including gestational age, available for 340 subjects, as a covariate did not alter the results pertaining to the effect of BW on cortical area within this cohort (n = 340, t(332) = 3.82, P = .0002). Adding height, available for 310 subjects, as a covariate changed the effects of BW on total surface area from t(303) = 2.82, P = .005 to t(302) = 2.28, P = .023. Within this smaller cohort (n = 310), the right anterior cingulate (P = .00036) was associated with BW after Bonferroni correction.
. | t-Value . | P-Value . | Adjusted P-Value . | B-Value (SE) . |
---|---|---|---|---|
Left hemisphere ROIs | ||||
Bank of superior temporal | 2.80 | .030 | 1 | 0.045 (0.021) |
Caudal anterior cingulate | 2.36 | .019 | 1 | 0.034 (0.015) |
Caudal middle frontal | 3.17 | .002 | .136 | 0.135 (0.043) |
Cuneus | 2.33 | .020 | 1 | 0.050 (0.021) |
Inferior parietal | 2.43 | .016 | 1 | 0.166 (0.068) |
Lateral orbitofrontal | 2.40 | .017 | 1 | 0.059 (0.025) |
Middle temporal | 4.38 | .0001 | .001 | 0.183 (0.042) |
Parahippocampal | 2.42 | .016 | 1 | 0.025 (0.010) |
Pars opercularis | 2.38 | .018 | 1 | 0.073 (0.031) |
Postcentral | 3.03 | .003 | .204 | 0.141 (0.048) |
Precentral | 2.35 | .019 | 1 | 0.110 (0.047) |
Precuneus | 2.56 | .011 | .748 | 0.110 (0.043) |
Rostral anterior cingulate | 3.81 | <.0001 | .011 | 0.047 (0.012) |
Rostral middle frontal | 1.98 | .048 | 1 | 0.158 (0.080) |
Superior frontal | 2.84 | .005 | .340 | 0.235 (0.083) |
Superior parietal | 3.09 | .002 | .136 | 0.191 (0.062) |
Supramarginal | 2.01 | .045 | 1 | 0.108 (0.054) |
Right hemisphere ROIs | ||||
Bank of superior temporal | 3.00 | .003 | .204 | 0.055 (0.018) |
Caudal middle frontal | 2.47 | .014 | .952 | 0.117 (0.047) |
Entorhinal | 3.62 | .001 | .068 | 0.024 (0.007) |
Inferior parietal | 3.64 | .0003 | .021 | 0.224 (0.067) |
Inferior temporal | 2.36 | .019 | 1 | 0.117 (0.050) |
Isthmus cingulate | 2.58 | .010 | .608 | 0.032 (0.013) |
Lateral orbitofrontal | 2.42 | .016 | 1 | 0.064 (0.026) |
Middle temporal | 3.03 | .003 | .204 | 0.126 (0.042) |
Parahippocampal | 3.83 | .0002 | .010 | 0.036 (0.009) |
Postcentral | 2.86 | .004 | .272 | 0.124 (0.043) |
Precentral | 3.30 | .001 | .068 | 0.165 (0.050) |
Rostral anterior cingulate | 3.99 | <.0001 | .005 | 0.045 (0.011) |
Rostral middle frontal | 3.00 | .003 | .204 | 0.221 (0.074) |
Superior frontal | 3.08 | .002 | .136 | 0.258 (0.084) |
. | t-Value . | P-Value . | Adjusted P-Value . | B-Value (SE) . |
---|---|---|---|---|
Left hemisphere ROIs | ||||
Bank of superior temporal | 2.80 | .030 | 1 | 0.045 (0.021) |
Caudal anterior cingulate | 2.36 | .019 | 1 | 0.034 (0.015) |
Caudal middle frontal | 3.17 | .002 | .136 | 0.135 (0.043) |
Cuneus | 2.33 | .020 | 1 | 0.050 (0.021) |
Inferior parietal | 2.43 | .016 | 1 | 0.166 (0.068) |
Lateral orbitofrontal | 2.40 | .017 | 1 | 0.059 (0.025) |
Middle temporal | 4.38 | .0001 | .001 | 0.183 (0.042) |
Parahippocampal | 2.42 | .016 | 1 | 0.025 (0.010) |
Pars opercularis | 2.38 | .018 | 1 | 0.073 (0.031) |
Postcentral | 3.03 | .003 | .204 | 0.141 (0.048) |
Precentral | 2.35 | .019 | 1 | 0.110 (0.047) |
Precuneus | 2.56 | .011 | .748 | 0.110 (0.043) |
Rostral anterior cingulate | 3.81 | <.0001 | .011 | 0.047 (0.012) |
Rostral middle frontal | 1.98 | .048 | 1 | 0.158 (0.080) |
Superior frontal | 2.84 | .005 | .340 | 0.235 (0.083) |
Superior parietal | 3.09 | .002 | .136 | 0.191 (0.062) |
Supramarginal | 2.01 | .045 | 1 | 0.108 (0.054) |
Right hemisphere ROIs | ||||
Bank of superior temporal | 3.00 | .003 | .204 | 0.055 (0.018) |
Caudal middle frontal | 2.47 | .014 | .952 | 0.117 (0.047) |
Entorhinal | 3.62 | .001 | .068 | 0.024 (0.007) |
Inferior parietal | 3.64 | .0003 | .021 | 0.224 (0.067) |
Inferior temporal | 2.36 | .019 | 1 | 0.117 (0.050) |
Isthmus cingulate | 2.58 | .010 | .608 | 0.032 (0.013) |
Lateral orbitofrontal | 2.42 | .016 | 1 | 0.064 (0.026) |
Middle temporal | 3.03 | .003 | .204 | 0.126 (0.042) |
Parahippocampal | 3.83 | .0002 | .010 | 0.036 (0.009) |
Postcentral | 2.86 | .004 | .272 | 0.124 (0.043) |
Precentral | 3.30 | .001 | .068 | 0.165 (0.050) |
Rostral anterior cingulate | 3.99 | <.0001 | .005 | 0.045 (0.011) |
Rostral middle frontal | 3.00 | .003 | .204 | 0.221 (0.074) |
Superior frontal | 3.08 | .002 | .136 | 0.258 (0.084) |
Note: Cortical area ROIs as dependent variables and birth weight (in grams), age, sex, and diagnosis as independent variables, n = 359. Only nominally significant values are shown. Bolded values are significant after Bonferroni adjustment for multiple comparisons (total number of ROIs = 68). B-values represent mm2 increase in cortical ROI area measure per increase in gram birth weight.
. | t-Value . | P-Value . | Adjusted P-Value . | B-Value (SE) . |
---|---|---|---|---|
Left hemisphere ROIs | ||||
Bank of superior temporal | 2.80 | .030 | 1 | 0.045 (0.021) |
Caudal anterior cingulate | 2.36 | .019 | 1 | 0.034 (0.015) |
Caudal middle frontal | 3.17 | .002 | .136 | 0.135 (0.043) |
Cuneus | 2.33 | .020 | 1 | 0.050 (0.021) |
Inferior parietal | 2.43 | .016 | 1 | 0.166 (0.068) |
Lateral orbitofrontal | 2.40 | .017 | 1 | 0.059 (0.025) |
Middle temporal | 4.38 | .0001 | .001 | 0.183 (0.042) |
Parahippocampal | 2.42 | .016 | 1 | 0.025 (0.010) |
Pars opercularis | 2.38 | .018 | 1 | 0.073 (0.031) |
Postcentral | 3.03 | .003 | .204 | 0.141 (0.048) |
Precentral | 2.35 | .019 | 1 | 0.110 (0.047) |
Precuneus | 2.56 | .011 | .748 | 0.110 (0.043) |
Rostral anterior cingulate | 3.81 | <.0001 | .011 | 0.047 (0.012) |
Rostral middle frontal | 1.98 | .048 | 1 | 0.158 (0.080) |
Superior frontal | 2.84 | .005 | .340 | 0.235 (0.083) |
Superior parietal | 3.09 | .002 | .136 | 0.191 (0.062) |
Supramarginal | 2.01 | .045 | 1 | 0.108 (0.054) |
Right hemisphere ROIs | ||||
Bank of superior temporal | 3.00 | .003 | .204 | 0.055 (0.018) |
Caudal middle frontal | 2.47 | .014 | .952 | 0.117 (0.047) |
Entorhinal | 3.62 | .001 | .068 | 0.024 (0.007) |
Inferior parietal | 3.64 | .0003 | .021 | 0.224 (0.067) |
Inferior temporal | 2.36 | .019 | 1 | 0.117 (0.050) |
Isthmus cingulate | 2.58 | .010 | .608 | 0.032 (0.013) |
Lateral orbitofrontal | 2.42 | .016 | 1 | 0.064 (0.026) |
Middle temporal | 3.03 | .003 | .204 | 0.126 (0.042) |
Parahippocampal | 3.83 | .0002 | .010 | 0.036 (0.009) |
Postcentral | 2.86 | .004 | .272 | 0.124 (0.043) |
Precentral | 3.30 | .001 | .068 | 0.165 (0.050) |
Rostral anterior cingulate | 3.99 | <.0001 | .005 | 0.045 (0.011) |
Rostral middle frontal | 3.00 | .003 | .204 | 0.221 (0.074) |
Superior frontal | 3.08 | .002 | .136 | 0.258 (0.084) |
. | t-Value . | P-Value . | Adjusted P-Value . | B-Value (SE) . |
---|---|---|---|---|
Left hemisphere ROIs | ||||
Bank of superior temporal | 2.80 | .030 | 1 | 0.045 (0.021) |
Caudal anterior cingulate | 2.36 | .019 | 1 | 0.034 (0.015) |
Caudal middle frontal | 3.17 | .002 | .136 | 0.135 (0.043) |
Cuneus | 2.33 | .020 | 1 | 0.050 (0.021) |
Inferior parietal | 2.43 | .016 | 1 | 0.166 (0.068) |
Lateral orbitofrontal | 2.40 | .017 | 1 | 0.059 (0.025) |
Middle temporal | 4.38 | .0001 | .001 | 0.183 (0.042) |
Parahippocampal | 2.42 | .016 | 1 | 0.025 (0.010) |
Pars opercularis | 2.38 | .018 | 1 | 0.073 (0.031) |
Postcentral | 3.03 | .003 | .204 | 0.141 (0.048) |
Precentral | 2.35 | .019 | 1 | 0.110 (0.047) |
Precuneus | 2.56 | .011 | .748 | 0.110 (0.043) |
Rostral anterior cingulate | 3.81 | <.0001 | .011 | 0.047 (0.012) |
Rostral middle frontal | 1.98 | .048 | 1 | 0.158 (0.080) |
Superior frontal | 2.84 | .005 | .340 | 0.235 (0.083) |
Superior parietal | 3.09 | .002 | .136 | 0.191 (0.062) |
Supramarginal | 2.01 | .045 | 1 | 0.108 (0.054) |
Right hemisphere ROIs | ||||
Bank of superior temporal | 3.00 | .003 | .204 | 0.055 (0.018) |
Caudal middle frontal | 2.47 | .014 | .952 | 0.117 (0.047) |
Entorhinal | 3.62 | .001 | .068 | 0.024 (0.007) |
Inferior parietal | 3.64 | .0003 | .021 | 0.224 (0.067) |
Inferior temporal | 2.36 | .019 | 1 | 0.117 (0.050) |
Isthmus cingulate | 2.58 | .010 | .608 | 0.032 (0.013) |
Lateral orbitofrontal | 2.42 | .016 | 1 | 0.064 (0.026) |
Middle temporal | 3.03 | .003 | .204 | 0.126 (0.042) |
Parahippocampal | 3.83 | .0002 | .010 | 0.036 (0.009) |
Postcentral | 2.86 | .004 | .272 | 0.124 (0.043) |
Precentral | 3.30 | .001 | .068 | 0.165 (0.050) |
Rostral anterior cingulate | 3.99 | <.0001 | .005 | 0.045 (0.011) |
Rostral middle frontal | 3.00 | .003 | .204 | 0.221 (0.074) |
Superior frontal | 3.08 | .002 | .136 | 0.258 (0.084) |
Note: Cortical area ROIs as dependent variables and birth weight (in grams), age, sex, and diagnosis as independent variables, n = 359. Only nominally significant values are shown. Bolded values are significant after Bonferroni adjustment for multiple comparisons (total number of ROIs = 68). B-values represent mm2 increase in cortical ROI area measure per increase in gram birth weight.
BW was not related to cortical thickness for any of the ROIs or mean thickness, and the cortical maps showed no significant effects of BW on cortical thickness (data not shown). No diagnosis-by-BW interaction effects were found, but total surface area and mean cortical thickness differed between the diagnostic groups (see online supplementary table 1). Cortical maps showed widespread prefrontal and temporal cortical thinning in schizophrenia patients compared with healthy controls (see online supplementary figure 2), in accordance with 2 previous reports from a cohort overlapping with this report.15,22
Working Memory Performance and IQ
In the whole-group analyses, smaller total cortical area was associated with poorer working memory performance (t(316) = 2.22, P = .027). BW was associated with working memory performance on a trend level (t(316) = 1.77, P = .078) in the whole-group analysis, but with opposite direction in the schizophrenia group (figure 3). Analyses stratified by diagnoses showed no associations of BW with working memory in schizophrenia patients but showed associations between lower BW (t(240) = 2.53, P = .012) and poorer working memory performance in the other groups. Although both BW and cortical area (t(240) = 2.50, P = .013) were related to working memory in the bipolar, other psychosis and healthy control groups, these effects did not meet the Baron-Kenny statistical criteria35 for cortical area to be a statistically significant mediator of BW on working memory performance because cortical area was only trend-level associated with working memory (t(239) = 1.90, P = .059) when BW was entered into the analysis. Smaller cortical area was associated with lower IQ estimates in the bipolar, other psychosis, and healthy control groups (t(255) = 2.49, P = .014) but not schizophrenia. BW was not associated with IQ in any of the groups.
Discussion
This study is, to our knowledge, the first to investigate effects of BW on cerebral cortical surface area and thickness in patients with psychotic disorders. We report novel associations between normal variation in BW and adult cortical surface area but not thickness, regardless of diagnosis. BW did not account for the differences in cortical morphology found between diagnostic groups. BW and cortical area were related to working memory performance.
Previous studies have shown associations between LBW and cortical area and thickness in healthy children36,37 and adolescents born with VLBW,9 and between normal birth weight variation and cortical area in healthy children and young adults.12 These findings expand on this by showing distinct cortical area effects of normal BW variation in psychotic patients and healthy controls. Indeed, these findings suggest that subtle, as well as severe, aberrances in prenatal neurodevelopment2,7,10 have lasting effects on cortical anatomy and that such effects are independent of a diagnosis of psychosis. One previous study reported a specific association between reduced cortical volumes and prenatal hypoxia in patients with schizophrenia, and suggested that prenatal adversities have larger effects on cortical anatomy in schizophrenia patients than in healthy controls.38 This study, however, provides no support for any specific effects of BW on cortical morphology in schizophrenia. This is consistent with our previous findings, from a different cohort, of a diagnosis-independent relationship between hypoxia-related OCs and decreased cortical folding in schizophrenia patients and healthy controls.39 If the smaller cortical area associated with lower BW interacts with, or adds to, other risk factors for developing psychosis,7 and thereby constitutes a more severe constellation in subjects who already are at genetic risk for psychotic disorders, remains unknown and is an important topic for future research.
We found no association between BW and cortical thickness. Accordingly, the cortical thickness reductions observed in patients with schizophrenia and bipolar disorder were not related to prenatal adversities as reflected by BW variation. This is in accordance with our earlier observations using composite OCs scores in a Swedish cohort.40 Other researchers have reported thinner cortices in subjects with VLBW.9 Such alterations may reflect more severe developmental trauma than what had occurred in this cohort, in which only 1 subject had VLBW. The fact that BW was related to cortical area, but not to cortical thickness, and that a diagnosis of psychosis was more strongly related to cortical thinning than to cortical area fits with the notion that cortical area and thickness are separate entities that follow distinct neurodevelopmental trajectories.14 Furthermore, it suggests that reduction of cortical area and thickness represent different pathophysiological mechanisms underlying psychotic disorders.
We found BW and cortical area to be independently related to working memory performance, rather than cortical area being a mediator of the effect of BW. The positive association between BW and working memory was present in all groups except in schizophrenia patients, in contrast to a study of a smaller cohort in which BW was associated with working memory in schizophrenia patients but not controls.19 It is biologically plausible that abnormalities in cortical circuitry underlying cognitive impairments have a prenatal origin because cortical neurons are generated, cortical pathways develop, and cortical synaptogenesis begins prenatally.41 BW was, however, not associated with IQ, in contrast to previous reports,17 including a study that used the same premorbid IQ estimate.18 IQ was significantly related to cortical area in all groups except in schizophrenia, but this effect appeared independent of BW.
Although we did not find an association between working memory and BW in schizophrenia patients, we found impaired working memory performance in this group, in accordance with the literature.21,42 Normal working memory performance involves dopaminergic input and requires balance in the prefrontal excitatory and inhibitory circuitry involving glutaminergic pyramidal neurons and GABAergic inhibitory interneurons.21 The lack of association between BW and working memory in schizophrenia patients suggest that other schizophrenia-specific factors, eg, abnormalities in dopamine and glutamate signaling, are of greater importance to working memory performance in schizophrenia than are cortical area reductions associated with BW.
These findings have implications for understanding brain development because they point toward a diagnosis-independent dose response, rather than dichotomous, association between prenatal adversities, and adult cortical anatomy and outcome, not only in LBW or psychotic subjects, as previously reported,7,19 but also in healthy and normal BW subjects. This is of importance to general obstetric health care, and in particular, the care for pregnant women with psychotic disorders or women with disadvantaged environmental living conditions.
Strengths of this study include the large subject sample with patients and control subjects from the same catchment area and the thorough clinical and neuropsychological assessment of each participating subject. MRI acquisition was performed on 1 scanner, with patients and control subjects scanned consecutively, and without any scanner upgrade during the study period. Data on BW were collected from the Medical Birth Registry of Norway; this avoids potential confounding by maternal recall bias.
Limitations include the uneven number of subjects within, and sex distribution between, the diagnostic groups although this was controlled for in the statistical analyses. Assessing the heritability of BW was beyond the scope of this study because the medical birth registry dates back to 1967 and does not include information on parental BW.
We conclude that cortical area appears to be a general, rather than a diagnosis-specific, marker of early neurodevelopment, with a dose-response association to normal BW variation. This is of importance to the understanding of brain pathology in psychotic disorders and normal brain development.
Funding
The Research Council of Norway (190311/V50, 167153/V50); and the South Eastern Norway Regional Health Authority (2008011, 2009037, 2011096).
Acknowledgments
We thank Merete Øibakken, Thomas Bjella, Martin Furan, and Eivind Bakken for technical assistance, and we acknowledge the services of the Medical Birth Registry of Norway.
IM has received Speaker’s honorarium from Janssen, AstraZeneca, and Lundbeck. OAA has received speaker’s honorarium from pharmaceutical companies AstraZeneca, BMS, GSK, Janssen-Cilag, Lundbeck. AMD is a founder, who holds equity in CorTechs Labs and also serves on the Scientific Advisory Board. The terms of this arrangement have been reviewed and approved by the University of California, San Diego, in accordance with its conflict of interest policies. Drs UKH, LMR, CR, CBH, EHL, AV, and IA report no conflict of interest.
References