- Email: [email protected]
Meta-analysis of adult height and birth length in schizophrenia
SCHRES-07516; No of Pages 5 Schizophrenia Research xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Schizophrenia Research journal homepage: www.elsevier.com/locate/schres
Meta-analysis of adult height and birth length in schizophrenia Keely Latham, Brian Kirkpatrick ⁎ Department of Psychiatry & Behavioral Sciences, University of Nevada, Reno School of Medicine, Reno, NV, United States
a r t i c l e
i n f o
Article history: Received 19 May 2017 Received in revised form 1 September 2017 Accepted 2 September 2017 Available online xxxx Keywords: Schizophrenia Height Birth length Meta-analysis Delayed effect Prenatal exposure
a b s t r a c t Objective: As a group, people with schizophrenia have a number of subtle anatomical abnormalities as well as physiological abnormalities that precede antipsychotic treatment. Some studies have also found shorter birth length or shorter adult height in people with schizophrenia compared to control subjects. We performed a systematic review and meta-analysis of studies of birth length and adult height in schizophrenia, following PRISMA guidelines (Prospero Registration # CRD42016043718). Data sources: We searched the PsycInfo, Web of Science, and PubMed databases for articles published 1947–2016. Study selection: Articles were included if they had data for patients diagnosed with schizophrenia and a matched control group of subjects without a psychotic disorder; both groups were measured for birth length and/or adult height (18 years or older); and the paper was published in English. Data extraction: One author extracted the data, which was verified by the other. Results: For adult height, six studies with 1,122 patients and 250,200 control subjects were included in analyses. There were six birth length studies, which included 984 patients and 976,296 controls. The patients did not differ from comparison subjects in birth length (effect size estimate = −0.03; CI: −0.09, 0.03), but adults were slightly shorter than comparison subjects (−0.15; −0.24, −0.06). In meta-regression of adult studies, the variables of first episode versus clinical sample, and population registry versus non-registry were not significant. Matching for several important variables was usually lacking in these studies. Conclusions: While there appears to be no difference in birth length between people with schizophrenia and comparison subjects, the former may be slighter shorter in adult life. The cause of such a discrepancy, if confirmed, is not clear, and lack of matching on potentially confounding variables undermines confidence in any conclusion. © 2017 Published by Elsevier B.V.
1. Introduction There is now an extensive literature on prenatal risk factors for adult disease, including cardiovascular disease, diabetes, and hypertension (Barker, 2003; Gluckman et al., 2007; Lawlor and Smith, 2005). A number of obstetric and prenatal problems are risk factors for schizophrenia (Cannon et al., 2002; Khandaker et al., 2013a, 2013b; Susser and St Clair, 2013; Knuesel et al., 2014). Consistent with that evidence, people with schizophrenia have an increased prevalence of minor physical anomalies, which are subtle anatomical abnormalities that have their origin in gestation (Franco et al., 2010; Xu et al., 2011). They also have several metabolic abnormalities, including abnormal glucose tolerance, increased inflammation, an elevated pulse pressure, increase prolactin, and decreased androgens in men, which are present prior to antipsychotic exposure (Miller et al., 2011; González-Blanco et al., 2016; Greenhalgh et al., 2017; Fernandez-Egea et al., 2009, 2011). These ⁎ Corresponding author at: Department of Psychiatry and Behavioral Sciences, University of Nevada, Reno School of Medicine, 5190 Neil Road, Suite 215, Reno, NV 89502, United States. E-mail address: [email protected] (B. Kirkpatrick).
abnormalities are consistent with the concept that for many patients, the pathophysiology of schizophrenia is related to difficulties in gestation. Prenatal events help determine adult height and birth length. While people with schizophrenia have been the subject of several studies of these variables, the results have not been consistent. The easily obtained measure of adult height would be a convenient measure that might be of theoretical interest for some areas of research. We performed a systematic review and meta-analysis of these studies to test the hypothesis that stature—adult height and birth length—is also abnormal in people with schizophrenia. 2. Materials and methods 2.1. Search strategy We followed PRISMA guidelines (Moher et al., 2009) and registered the meta-analysis through PROSPERO (https://www.crd.york.ac.uk/ PROSPERO/) as “Height in Schizophrenia Spectrum Disorders,” registration number CRD42016043718. The literature search included articles from 1947 to 2016 on PsycInfo, Web of Science, and PubMed databases,
https://doi.org/10.1016/j.schres.2017.09.002 0920-9964/© 2017 Published by Elsevier B.V.
Please cite this article as: Latham, K., Kirkpatrick, B., Meta-analysis of adult height and birth length in schizophrenia, Schizophr. Res. (2017), https://doi.org/10.1016/j.schres.2017.09.002
2
K. Latham, B. Kirkpatrick / Schizophrenia Research xxx (2017) xxx–xxx
using search terms of “schizophrenia” crossed with “body height”. The bibliographies of the articles found by computer search were then examined for other relevant articles. Both authors (KL and BK) examined the titles and abstracts to find possibly relevant articles by both authors; they then inspected these articles to determine which fulfilled the inclusion and exclusion criteria. 2.2. Inclusion/exclusion criteria To meet inclusion criteria a study had to include data for patients diagnosed with schizophrenia or schizoaffective disorder and a matched control group of subjects without a psychotic disorder; these two groups were measured for birth length, adult height (18 years or older), or both; and the paper was published in English. 2.3. Data analysis One author (KL) recorded the data from the studies, which was verified by the second (BK). All analyses were conducted separately for
birth length and adult height and were performed using Stata 13.1 software program. Effect size (ES) estimates using Hedges' g were calculated separately for adult height and birth length. This statistic quantifies the size of the difference between group means relative to their standard deviations (http://www.statisticshowto.com/hedges-g/). Random effects, pooled ES estimates and 95% confidence intervals were calculated using the method of DerSimonian and Laird (1986); p-values were considered statistically significant at the p b 0.05 level. We also examined forest and funnel plots for adult height and birth length. For adult height, meta-regression was conducted for the variables of 1) first-episode patients (2 studies) vs. a clinic sample (4 studies), and 2) the use of registry (2 studies) vs. non-registry subjects (4 studies). We also compared the ES estimate of 1) the two studies (one of adult height, one of birth length; respectively Sugawara et al., 2012 and Perrin et al., 2007) that included schizoaffective disorder patients to the ES estimate of the other five adult height or birth length studies, and 2) the two birth length studies (Haukka et al., 2008; Woerner et al., 1971) that used siblings as control subjects, to the other four birth length studies.
Fig. 1. Selection of studies for meta-analysis.
Please cite this article as: Latham, K., Kirkpatrick, B., Meta-analysis of adult height and birth length in schizophrenia, Schizophr. Res. (2017), https://doi.org/10.1016/j.schres.2017.09.002
K. Latham, B. Kirkpatrick / Schizophrenia Research xxx (2017) xxx–xxx
3
Fig. 2. Forest plot adult height. X-axis is effect size estimate (Hedges' g). Negative values represent shorter adult height in people with schizophrenia compared to control subjects.
3. Results The computerized search, conducted in November 2016, yielded 308 articles. Using the selection approach outlined above these were narrowed down to 10 that were included in analyses (see Fig. 1 for details of inclusion and exclusion of the articles). Of these 10 articles, 4 had data on adult height but not birth length (Houston and Bloom, 1975; Nopoulos et al., 1998; Singer et al., 1972; Sugawara et al., 2012), and another 4 had data on birth length but not adult height (Gunnell et al., 2005; Haukka et al., 2008; Perrin et al., 2007; Woerner et al., 1971). Two articles (Gunnell et al., 2003; Sorensen et al., 2006) provided data for both adult height and birth length and were used in both analyses. One article (Nopoulos et al., 1998) gave two different values for the number of control subjects; in the forest plots and in our primary
analysis we used the smaller number given in the methods as we were not able to confirm the number in the text with the first author. These two numbers differed by just 20 control subjects (out of 250,200), and using the other number provided made very little difference in ES estimate (data not shown). 3.1. Adult height analysis The articles for adult height included 1122 patients and 250,200 controls. The patients were significantly shorter than the control subjects (ES estimate = − 0.15; 95% CI −0.24, − 0.06; Fig. 2). The funnel plot for adult height (Fig. 3) did not show obvious publication bias but the number of studies was small. Heterogeneity was not found to be significant for studies included in adult height [I-squared = 0.9%, p = 0.410].
Fig. 3. Funnel plot adult height. X-axis is effect size estimate (Hedges' g). Negative values represent shorter adult height in people with schizophrenia compared to control subjects.
Please cite this article as: Latham, K., Kirkpatrick, B., Meta-analysis of adult height and birth length in schizophrenia, Schizophr. Res. (2017), https://doi.org/10.1016/j.schres.2017.09.002
4
K. Latham, B. Kirkpatrick / Schizophrenia Research xxx (2017) xxx–xxx
In meta-regression, there was no significant effect for first episode vs. non-first episode patient groups (t = 1.17, p = 0.31), or for registry vs. non-registry comparison groups (t = −0.64, p = 0.56; Supplementary Figs. S1 & S2). The ES of the one study that included schizoaffective subjects (Sugawara et al., 2012) was 0.00; the ES estimate of the other five studies combined was −0.19. Separate data for males and females is available in only two studies of adult height. The forest plots for these three studies, with males and females separated, can be found in the supplementary materials.
4. Discussion
There were some significant weaknesses in our analyses. While the number of subjects was large for both birth length and adult height, the number of studies was small. The matching between the patient and comparison groups in these studies was usually limited to gender for height and birth length, and to age and gender for adult height. Gestational age was not always considered for birth length. Although the groups were matched for gender, for this variable gender-specific data was usually not provided available; as a consequence it was not possible to assess any male/female differences thoroughly. Details on ethnicity and socioeconomic status of the family of origin were also lacking. A potentially confounding variable is whether the patients and controls were equally upright—that is, maintaining erect posture—while measured, although it seems likely the investigators would have taken care with this aspect of measuring adult height. A final ambiguity is whether the differences might have been due to subtle lordosis or kyphosis, which might have affected the measurement of length and height. These weaknesses undermine confidence in the finding of a difference in adult height. Should the results of our meta-analysis accurately reflect adult height and birth length in schizophrenia despite these limitations, why there would be the discrepancy we found in these two measures is not clear. This failure to replicate an effect across these two study designs weakens one's confidence that there is a difference in adult height between people with schizophrenia and the general population, as does the small ES for height. The small ES for adult height further suggests that such a difference—even when compared to siblings, which should increase the signal to noise ratio—may not often be a variable of use in future research. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.schres.2017.09.002.
In a meta-analysis, we found a significant difference between people with schizophrenia and comparison subjects in adult height, but not in birth length. In meta-regression, the variables of first episode vs. clinical patient groups, and registry vs/non-registry control groups did not appear to influence ES estimate. The number of studies was not large enough to provide a robust evaluation of publication bias. Although the number of studies was small, the differences between the effect sizes of the studies that included schizoaffective subjects raise the possibility that the height difference does not extend to that group.
Conflict of interest Dr. Kirkpatrick receives licensing royalties from ProPhase LLC for use of the Brief Negative Symptom Scale (BNSS) by for-profit groups; these fees are donated to the Brain and Behavior Research Foundation. He has also received honoraria and travel support from ProPhase LLC for training pharmaceutical company raters on the BNSS, consulting fees and travel support from Genentech/Roche, Minerva Neurosciences, and ProPhase LLC, consulting fees from anonymized pharmaceutical companies through Decision Resources, Inc. and from an investment capital company through Guideposts. Dr. Kirkpatrick also receives fees from Walsh Medical Media for editorial services, and received fees for editorial services from Physicians Postgraduate Press, Inc.
3.2. Birth length analysis The articles for birth length included 984 patients and 976,296 controls. The two groups did not differ significantly in ES estimate (−0.03; 95% CI −0.09, 0.03; Fig. 4). The funnel plot for birth length (Fig. 5) did not show obvious publication bias but the number of studies was small. Heterogeneity was not found to be significant for studies included in birth length [I-squared = 0.0%, p = 0.932]. The ES estimate of the two studies that included siblings (Haukka et al., 2008; Woerner et al., 1971) was 0.12; the ES estimate for the other four studies was − 0.03. The ES of the one study that included schizoaffective subjects (Perrin et al., 2007) was 0.03; the ES estimate of the other five studies combined was −0.03. Separate data for males and females is available in only one study of adult height. The forest plots for that study, with males and females separated, can be found in the supplementary materials.
Fig. 4. Forest plot birth length. X-axis is effect size estimate (Hedges' g). Negative values represent shorter birth length in people with schizophrenia compared to control subjects.
Please cite this article as: Latham, K., Kirkpatrick, B., Meta-analysis of adult height and birth length in schizophrenia, Schizophr. Res. (2017), https://doi.org/10.1016/j.schres.2017.09.002
K. Latham, B. Kirkpatrick / Schizophrenia Research xxx (2017) xxx–xxx
5
Fig. 5. Funnel plot birth length. X-axis is effect size estimate (Hedges' g). Negative values represent shorter birth length in people with schizophrenia compared to control subjects.
Ms. Latham has no interests to declare. Contributors Keely Latham: Keely Latham executed the data analysis, helped select articles and extracted data as described in the manuscript, and collaborated in writing the manuscript. Dr. Brian Kirkpatrick: Dr. Brian Kirkpatrick designed the study, wrote the protocol, helped select articles and checked extracted data as described in the manuscript, and collaborated in writing the paper. **All authors contributed to and have approved the final manuscript. Role of funding source Not applicable. Acknowledgment No acknowledgements.
References Barker, D.J., 2003. The developmental origins of adult disease. Eur. J. Epidemiol. 18 (8), 733–736. Cannon, M., Jones, P.B., Murray, R.M., 2002. Obstetric complications and schizophrenia: historical and meta-analytic review. Am. J. Psychiatry 159 (7), 1080–1092. DerSimonian, R., Laird, N., 1986. Meta-analysis in clinical trials. Control. Clin. Trials 7 (3), 177–188. Fernandez-Egea, E., Bernardo, M., Heaphy, C.M., Griffith, J.K., Parellada, E., Esmatjes, E., Conget, I., Nguyen, L., George, V., Stoppler, H., Kirkpatrick, B., 2009. Telomere length and pulse pressure in newly diagnosed, antipsychotic-naive patients with nonaffective psychosis. Schizophr. Bull. 35 (2), 437–442. Fernandez-Egea, E., Garcia-Rizo, C., Miller, B., Parellada, E., Justicia, A., Bernardo, M., Kirkpatrick, B., 2011. Testosterone in newly diagnosed, antipsychotic-naive men with nonaffective psychosis: a test of the accelerated aging hypothesis. Psychosom. Med. 73 (8), 643–647. Franco, J.G., Valero, J., Labad, A., 2010. Minor physical anomalies and schizophrenia: literature review. Actas Esp. Psiquiatr. 38 (6), 365–371. Gluckman, P.D., Lillycrop, K.A., Vickers, M.H., Pleasants, A.B., Phillips, E.S., Beedle, A.S., Burdge, G.C., Hanson, M.A., 2007. Metabolic plasticity during mammalian development is directionally dependent on early nutritional status. Proc. Natl. Acad. Sci. U. S. A. 104 (31), 12796–12800. González-Blanco, L., Greenhalgh, A.M., Garcia-Rizo, C., Fernandez-Egea, E., Miller, B.J., Kirkpatrick, B., 2016. Prolactin concentrations in antipsychotic-naïve patients with schizophrenia and related disorders: a meta-analysis. Schizophr. Res. 174 (1–3), 156–160. Greenhalgh, A.M., González-Blanco, L., Garcia-Rizo, C., et al., 2017. Meta-analysis of glucose tolerance, insulin, and insulin resistance in antipsychotic-naïve patients with nonaffective psychosis. Schizophr. Res. 179, 57–63.
Gunnell, D., Rasmussen, F., Fouskakis, D., Tynelius, P., Harrison, G., 2003. Patterns of fetal and childhood growth and the development of psychosis in young males: a cohort study. Am. J. Epidemiol. 158, 291–300. Gunnell, D., Harrison, G., Whitley, E., Lewis, G., Tynelius, P., Rasmussen, F., 2005. The association of fetal and childhood growth with risk of schizophrenia. Cohort study of 720,000 Swedish men and women. Schizophr. Res. 79, 315–322. Haukka, J., Suvisaari, J., Häkkinen, L., Lönnqvist, J., 2008. Growth pattern and risk of schizophrenia. Psychol. Med. 38 (1), 63–70. Houston, B.K., Bloom, L.J., 1975. Constitutional factors, stature, and chronic schizophrenia. J. Clin. Psychol. 31 (1), 26–29. Khandaker, G.M., Zimbron, J., Lewis, G., Jones, P.B., 2013a. Prenatal maternal infection, neurodevelopment and adult schizophrenia: a systematic review of populationbased studies. Psychol. Med. 43 (2), 239–257. Khandaker, G.M., Zimbron, J., Lewis, G., Jones, P.B., 2013b. Prenatal maternal infection, neurodevelopment and adult schizophrenia: a systematic review of populationbased studies. Psychol. Med. 43 (2), 239–257. Knuesel, I., Chicha, L., Britschgi, M., Schobel, S.A., Bodmer, M., Hellings, J.A., Toovey, S., Prinssen, E.P., 2014. Maternal immune activation and abnormal brain development across CNS disorders. Nature reviews. Neurology 10 (11), 643–660. Lawlor, D.A., Smith, G.D., 2005. Early life determinants of adult blood pressure. Curr. Opin. Nephrol. Hypertens. 14 (3), 259–264. Miller, B.J., Buckley, P., Seabolt, W., Mellor, A., Kirkpatrick, B., 2011. Meta-analysis of cytokine alterations in schizophrenia: clinical status and antipsychotic effects. Biol. Psychiatry 70 (7), 663–671. Moher, D., Liberati, A., Tetzlaff, J., Altman, D.G., The PRISMA Group, 2009. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 6 (7), e1000097. https://doi.org/10.1371/journal.pmed.1000097. Nopoulos, P., Flaum, M., Arndt, S., Andreasen, N., 1998. Morphometry in schizophrenia revisited: height and its relationship to pre-morbid function. Psychol. Med. 28, 655–663. Perrin, M.A., Chen, H., Sandberg, D.E., Malaspina, D., Brown, A.S., 2007. Growth trajectory during early life and risk of adult schizophrenia. Br. J. Psychiatry 191, 512–520. Singer, K., Chang, P.T., Hsu, G.L., 1972. Physique, personality and mental illness in the Southern Chinese. Br. J. Psychiatry 121, 315–319. Sorensen, H.J., Mortensen, E.L., Reinisch, J.M., Mednick, S.A., 2006. Height, weight and body mass index in early adulthood and risk of schizophrenia. Acta Psychiatr. Scand. 114 (1), 49–54. Sugawara, N., Yasui-Furukori, N., Tsuchimine, S., et al., 2012. Body composition in patients with schizophrenia: comparison with healthy controls. Ann. General Psychiatry 11. https://doi.org/10.1186/1744-859X-11-11. Susser, E., St Clair, D., 2013. Prenatal famine and adult mental illness: interpreting concordant and discordant results from the Dutch and Chinese famines. Soc. Sci. Med. 97, 325–330. Woerner, M.G., Pollack, M., Klein, D.F., 1971. Birth weight and length in schizophrenics personality disorders and their siblings. Br. J. Psychiatry 118 (545), 461–464. Xu, T., Chan, R.C., Compton, M.T., 2011. Minor physical anomalies in patients with schizophrenia, unaffected first-degree relatives, and healthy controls: a meta-analysis. PLoS One 6 (9), e24129.
Please cite this article as: Latham, K., Kirkpatrick, B., Meta-analysis of adult height and birth length in schizophrenia, Schizophr. Res. (2017), https://doi.org/10.1016/j.schres.2017.09.002
Contents lists available at ScienceDirect
Schizophrenia Research journal homepage: www.elsevier.com/locate/schres
Meta-analysis of adult height and birth length in schizophrenia Keely Latham, Brian Kirkpatrick ⁎ Department of Psychiatry & Behavioral Sciences, University of Nevada, Reno School of Medicine, Reno, NV, United States
a r t i c l e
i n f o
Article history: Received 19 May 2017 Received in revised form 1 September 2017 Accepted 2 September 2017 Available online xxxx Keywords: Schizophrenia Height Birth length Meta-analysis Delayed effect Prenatal exposure
a b s t r a c t Objective: As a group, people with schizophrenia have a number of subtle anatomical abnormalities as well as physiological abnormalities that precede antipsychotic treatment. Some studies have also found shorter birth length or shorter adult height in people with schizophrenia compared to control subjects. We performed a systematic review and meta-analysis of studies of birth length and adult height in schizophrenia, following PRISMA guidelines (Prospero Registration # CRD42016043718). Data sources: We searched the PsycInfo, Web of Science, and PubMed databases for articles published 1947–2016. Study selection: Articles were included if they had data for patients diagnosed with schizophrenia and a matched control group of subjects without a psychotic disorder; both groups were measured for birth length and/or adult height (18 years or older); and the paper was published in English. Data extraction: One author extracted the data, which was verified by the other. Results: For adult height, six studies with 1,122 patients and 250,200 control subjects were included in analyses. There were six birth length studies, which included 984 patients and 976,296 controls. The patients did not differ from comparison subjects in birth length (effect size estimate = −0.03; CI: −0.09, 0.03), but adults were slightly shorter than comparison subjects (−0.15; −0.24, −0.06). In meta-regression of adult studies, the variables of first episode versus clinical sample, and population registry versus non-registry were not significant. Matching for several important variables was usually lacking in these studies. Conclusions: While there appears to be no difference in birth length between people with schizophrenia and comparison subjects, the former may be slighter shorter in adult life. The cause of such a discrepancy, if confirmed, is not clear, and lack of matching on potentially confounding variables undermines confidence in any conclusion. © 2017 Published by Elsevier B.V.
1. Introduction There is now an extensive literature on prenatal risk factors for adult disease, including cardiovascular disease, diabetes, and hypertension (Barker, 2003; Gluckman et al., 2007; Lawlor and Smith, 2005). A number of obstetric and prenatal problems are risk factors for schizophrenia (Cannon et al., 2002; Khandaker et al., 2013a, 2013b; Susser and St Clair, 2013; Knuesel et al., 2014). Consistent with that evidence, people with schizophrenia have an increased prevalence of minor physical anomalies, which are subtle anatomical abnormalities that have their origin in gestation (Franco et al., 2010; Xu et al., 2011). They also have several metabolic abnormalities, including abnormal glucose tolerance, increased inflammation, an elevated pulse pressure, increase prolactin, and decreased androgens in men, which are present prior to antipsychotic exposure (Miller et al., 2011; González-Blanco et al., 2016; Greenhalgh et al., 2017; Fernandez-Egea et al., 2009, 2011). These ⁎ Corresponding author at: Department of Psychiatry and Behavioral Sciences, University of Nevada, Reno School of Medicine, 5190 Neil Road, Suite 215, Reno, NV 89502, United States. E-mail address: [email protected] (B. Kirkpatrick).
abnormalities are consistent with the concept that for many patients, the pathophysiology of schizophrenia is related to difficulties in gestation. Prenatal events help determine adult height and birth length. While people with schizophrenia have been the subject of several studies of these variables, the results have not been consistent. The easily obtained measure of adult height would be a convenient measure that might be of theoretical interest for some areas of research. We performed a systematic review and meta-analysis of these studies to test the hypothesis that stature—adult height and birth length—is also abnormal in people with schizophrenia. 2. Materials and methods 2.1. Search strategy We followed PRISMA guidelines (Moher et al., 2009) and registered the meta-analysis through PROSPERO (https://www.crd.york.ac.uk/ PROSPERO/) as “Height in Schizophrenia Spectrum Disorders,” registration number CRD42016043718. The literature search included articles from 1947 to 2016 on PsycInfo, Web of Science, and PubMed databases,
https://doi.org/10.1016/j.schres.2017.09.002 0920-9964/© 2017 Published by Elsevier B.V.
Please cite this article as: Latham, K., Kirkpatrick, B., Meta-analysis of adult height and birth length in schizophrenia, Schizophr. Res. (2017), https://doi.org/10.1016/j.schres.2017.09.002
2
K. Latham, B. Kirkpatrick / Schizophrenia Research xxx (2017) xxx–xxx
using search terms of “schizophrenia” crossed with “body height”. The bibliographies of the articles found by computer search were then examined for other relevant articles. Both authors (KL and BK) examined the titles and abstracts to find possibly relevant articles by both authors; they then inspected these articles to determine which fulfilled the inclusion and exclusion criteria. 2.2. Inclusion/exclusion criteria To meet inclusion criteria a study had to include data for patients diagnosed with schizophrenia or schizoaffective disorder and a matched control group of subjects without a psychotic disorder; these two groups were measured for birth length, adult height (18 years or older), or both; and the paper was published in English. 2.3. Data analysis One author (KL) recorded the data from the studies, which was verified by the second (BK). All analyses were conducted separately for
birth length and adult height and were performed using Stata 13.1 software program. Effect size (ES) estimates using Hedges' g were calculated separately for adult height and birth length. This statistic quantifies the size of the difference between group means relative to their standard deviations (http://www.statisticshowto.com/hedges-g/). Random effects, pooled ES estimates and 95% confidence intervals were calculated using the method of DerSimonian and Laird (1986); p-values were considered statistically significant at the p b 0.05 level. We also examined forest and funnel plots for adult height and birth length. For adult height, meta-regression was conducted for the variables of 1) first-episode patients (2 studies) vs. a clinic sample (4 studies), and 2) the use of registry (2 studies) vs. non-registry subjects (4 studies). We also compared the ES estimate of 1) the two studies (one of adult height, one of birth length; respectively Sugawara et al., 2012 and Perrin et al., 2007) that included schizoaffective disorder patients to the ES estimate of the other five adult height or birth length studies, and 2) the two birth length studies (Haukka et al., 2008; Woerner et al., 1971) that used siblings as control subjects, to the other four birth length studies.
Fig. 1. Selection of studies for meta-analysis.
Please cite this article as: Latham, K., Kirkpatrick, B., Meta-analysis of adult height and birth length in schizophrenia, Schizophr. Res. (2017), https://doi.org/10.1016/j.schres.2017.09.002
K. Latham, B. Kirkpatrick / Schizophrenia Research xxx (2017) xxx–xxx
3
Fig. 2. Forest plot adult height. X-axis is effect size estimate (Hedges' g). Negative values represent shorter adult height in people with schizophrenia compared to control subjects.
3. Results The computerized search, conducted in November 2016, yielded 308 articles. Using the selection approach outlined above these were narrowed down to 10 that were included in analyses (see Fig. 1 for details of inclusion and exclusion of the articles). Of these 10 articles, 4 had data on adult height but not birth length (Houston and Bloom, 1975; Nopoulos et al., 1998; Singer et al., 1972; Sugawara et al., 2012), and another 4 had data on birth length but not adult height (Gunnell et al., 2005; Haukka et al., 2008; Perrin et al., 2007; Woerner et al., 1971). Two articles (Gunnell et al., 2003; Sorensen et al., 2006) provided data for both adult height and birth length and were used in both analyses. One article (Nopoulos et al., 1998) gave two different values for the number of control subjects; in the forest plots and in our primary
analysis we used the smaller number given in the methods as we were not able to confirm the number in the text with the first author. These two numbers differed by just 20 control subjects (out of 250,200), and using the other number provided made very little difference in ES estimate (data not shown). 3.1. Adult height analysis The articles for adult height included 1122 patients and 250,200 controls. The patients were significantly shorter than the control subjects (ES estimate = − 0.15; 95% CI −0.24, − 0.06; Fig. 2). The funnel plot for adult height (Fig. 3) did not show obvious publication bias but the number of studies was small. Heterogeneity was not found to be significant for studies included in adult height [I-squared = 0.9%, p = 0.410].
Fig. 3. Funnel plot adult height. X-axis is effect size estimate (Hedges' g). Negative values represent shorter adult height in people with schizophrenia compared to control subjects.
Please cite this article as: Latham, K., Kirkpatrick, B., Meta-analysis of adult height and birth length in schizophrenia, Schizophr. Res. (2017), https://doi.org/10.1016/j.schres.2017.09.002
4
K. Latham, B. Kirkpatrick / Schizophrenia Research xxx (2017) xxx–xxx
In meta-regression, there was no significant effect for first episode vs. non-first episode patient groups (t = 1.17, p = 0.31), or for registry vs. non-registry comparison groups (t = −0.64, p = 0.56; Supplementary Figs. S1 & S2). The ES of the one study that included schizoaffective subjects (Sugawara et al., 2012) was 0.00; the ES estimate of the other five studies combined was −0.19. Separate data for males and females is available in only two studies of adult height. The forest plots for these three studies, with males and females separated, can be found in the supplementary materials.
4. Discussion
There were some significant weaknesses in our analyses. While the number of subjects was large for both birth length and adult height, the number of studies was small. The matching between the patient and comparison groups in these studies was usually limited to gender for height and birth length, and to age and gender for adult height. Gestational age was not always considered for birth length. Although the groups were matched for gender, for this variable gender-specific data was usually not provided available; as a consequence it was not possible to assess any male/female differences thoroughly. Details on ethnicity and socioeconomic status of the family of origin were also lacking. A potentially confounding variable is whether the patients and controls were equally upright—that is, maintaining erect posture—while measured, although it seems likely the investigators would have taken care with this aspect of measuring adult height. A final ambiguity is whether the differences might have been due to subtle lordosis or kyphosis, which might have affected the measurement of length and height. These weaknesses undermine confidence in the finding of a difference in adult height. Should the results of our meta-analysis accurately reflect adult height and birth length in schizophrenia despite these limitations, why there would be the discrepancy we found in these two measures is not clear. This failure to replicate an effect across these two study designs weakens one's confidence that there is a difference in adult height between people with schizophrenia and the general population, as does the small ES for height. The small ES for adult height further suggests that such a difference—even when compared to siblings, which should increase the signal to noise ratio—may not often be a variable of use in future research. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.schres.2017.09.002.
In a meta-analysis, we found a significant difference between people with schizophrenia and comparison subjects in adult height, but not in birth length. In meta-regression, the variables of first episode vs. clinical patient groups, and registry vs/non-registry control groups did not appear to influence ES estimate. The number of studies was not large enough to provide a robust evaluation of publication bias. Although the number of studies was small, the differences between the effect sizes of the studies that included schizoaffective subjects raise the possibility that the height difference does not extend to that group.
Conflict of interest Dr. Kirkpatrick receives licensing royalties from ProPhase LLC for use of the Brief Negative Symptom Scale (BNSS) by for-profit groups; these fees are donated to the Brain and Behavior Research Foundation. He has also received honoraria and travel support from ProPhase LLC for training pharmaceutical company raters on the BNSS, consulting fees and travel support from Genentech/Roche, Minerva Neurosciences, and ProPhase LLC, consulting fees from anonymized pharmaceutical companies through Decision Resources, Inc. and from an investment capital company through Guideposts. Dr. Kirkpatrick also receives fees from Walsh Medical Media for editorial services, and received fees for editorial services from Physicians Postgraduate Press, Inc.
3.2. Birth length analysis The articles for birth length included 984 patients and 976,296 controls. The two groups did not differ significantly in ES estimate (−0.03; 95% CI −0.09, 0.03; Fig. 4). The funnel plot for birth length (Fig. 5) did not show obvious publication bias but the number of studies was small. Heterogeneity was not found to be significant for studies included in birth length [I-squared = 0.0%, p = 0.932]. The ES estimate of the two studies that included siblings (Haukka et al., 2008; Woerner et al., 1971) was 0.12; the ES estimate for the other four studies was − 0.03. The ES of the one study that included schizoaffective subjects (Perrin et al., 2007) was 0.03; the ES estimate of the other five studies combined was −0.03. Separate data for males and females is available in only one study of adult height. The forest plots for that study, with males and females separated, can be found in the supplementary materials.
Fig. 4. Forest plot birth length. X-axis is effect size estimate (Hedges' g). Negative values represent shorter birth length in people with schizophrenia compared to control subjects.
Please cite this article as: Latham, K., Kirkpatrick, B., Meta-analysis of adult height and birth length in schizophrenia, Schizophr. Res. (2017), https://doi.org/10.1016/j.schres.2017.09.002
K. Latham, B. Kirkpatrick / Schizophrenia Research xxx (2017) xxx–xxx
5
Fig. 5. Funnel plot birth length. X-axis is effect size estimate (Hedges' g). Negative values represent shorter birth length in people with schizophrenia compared to control subjects.
Ms. Latham has no interests to declare. Contributors Keely Latham: Keely Latham executed the data analysis, helped select articles and extracted data as described in the manuscript, and collaborated in writing the manuscript. Dr. Brian Kirkpatrick: Dr. Brian Kirkpatrick designed the study, wrote the protocol, helped select articles and checked extracted data as described in the manuscript, and collaborated in writing the paper. **All authors contributed to and have approved the final manuscript. Role of funding source Not applicable. Acknowledgment No acknowledgements.
References Barker, D.J., 2003. The developmental origins of adult disease. Eur. J. Epidemiol. 18 (8), 733–736. Cannon, M., Jones, P.B., Murray, R.M., 2002. Obstetric complications and schizophrenia: historical and meta-analytic review. Am. J. Psychiatry 159 (7), 1080–1092. DerSimonian, R., Laird, N., 1986. Meta-analysis in clinical trials. Control. Clin. Trials 7 (3), 177–188. Fernandez-Egea, E., Bernardo, M., Heaphy, C.M., Griffith, J.K., Parellada, E., Esmatjes, E., Conget, I., Nguyen, L., George, V., Stoppler, H., Kirkpatrick, B., 2009. Telomere length and pulse pressure in newly diagnosed, antipsychotic-naive patients with nonaffective psychosis. Schizophr. Bull. 35 (2), 437–442. Fernandez-Egea, E., Garcia-Rizo, C., Miller, B., Parellada, E., Justicia, A., Bernardo, M., Kirkpatrick, B., 2011. Testosterone in newly diagnosed, antipsychotic-naive men with nonaffective psychosis: a test of the accelerated aging hypothesis. Psychosom. Med. 73 (8), 643–647. Franco, J.G., Valero, J., Labad, A., 2010. Minor physical anomalies and schizophrenia: literature review. Actas Esp. Psiquiatr. 38 (6), 365–371. Gluckman, P.D., Lillycrop, K.A., Vickers, M.H., Pleasants, A.B., Phillips, E.S., Beedle, A.S., Burdge, G.C., Hanson, M.A., 2007. Metabolic plasticity during mammalian development is directionally dependent on early nutritional status. Proc. Natl. Acad. Sci. U. S. A. 104 (31), 12796–12800. González-Blanco, L., Greenhalgh, A.M., Garcia-Rizo, C., Fernandez-Egea, E., Miller, B.J., Kirkpatrick, B., 2016. Prolactin concentrations in antipsychotic-naïve patients with schizophrenia and related disorders: a meta-analysis. Schizophr. Res. 174 (1–3), 156–160. Greenhalgh, A.M., González-Blanco, L., Garcia-Rizo, C., et al., 2017. Meta-analysis of glucose tolerance, insulin, and insulin resistance in antipsychotic-naïve patients with nonaffective psychosis. Schizophr. Res. 179, 57–63.
Gunnell, D., Rasmussen, F., Fouskakis, D., Tynelius, P., Harrison, G., 2003. Patterns of fetal and childhood growth and the development of psychosis in young males: a cohort study. Am. J. Epidemiol. 158, 291–300. Gunnell, D., Harrison, G., Whitley, E., Lewis, G., Tynelius, P., Rasmussen, F., 2005. The association of fetal and childhood growth with risk of schizophrenia. Cohort study of 720,000 Swedish men and women. Schizophr. Res. 79, 315–322. Haukka, J., Suvisaari, J., Häkkinen, L., Lönnqvist, J., 2008. Growth pattern and risk of schizophrenia. Psychol. Med. 38 (1), 63–70. Houston, B.K., Bloom, L.J., 1975. Constitutional factors, stature, and chronic schizophrenia. J. Clin. Psychol. 31 (1), 26–29. Khandaker, G.M., Zimbron, J., Lewis, G., Jones, P.B., 2013a. Prenatal maternal infection, neurodevelopment and adult schizophrenia: a systematic review of populationbased studies. Psychol. Med. 43 (2), 239–257. Khandaker, G.M., Zimbron, J., Lewis, G., Jones, P.B., 2013b. Prenatal maternal infection, neurodevelopment and adult schizophrenia: a systematic review of populationbased studies. Psychol. Med. 43 (2), 239–257. Knuesel, I., Chicha, L., Britschgi, M., Schobel, S.A., Bodmer, M., Hellings, J.A., Toovey, S., Prinssen, E.P., 2014. Maternal immune activation and abnormal brain development across CNS disorders. Nature reviews. Neurology 10 (11), 643–660. Lawlor, D.A., Smith, G.D., 2005. Early life determinants of adult blood pressure. Curr. Opin. Nephrol. Hypertens. 14 (3), 259–264. Miller, B.J., Buckley, P., Seabolt, W., Mellor, A., Kirkpatrick, B., 2011. Meta-analysis of cytokine alterations in schizophrenia: clinical status and antipsychotic effects. Biol. Psychiatry 70 (7), 663–671. Moher, D., Liberati, A., Tetzlaff, J., Altman, D.G., The PRISMA Group, 2009. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 6 (7), e1000097. https://doi.org/10.1371/journal.pmed.1000097. Nopoulos, P., Flaum, M., Arndt, S., Andreasen, N., 1998. Morphometry in schizophrenia revisited: height and its relationship to pre-morbid function. Psychol. Med. 28, 655–663. Perrin, M.A., Chen, H., Sandberg, D.E., Malaspina, D., Brown, A.S., 2007. Growth trajectory during early life and risk of adult schizophrenia. Br. J. Psychiatry 191, 512–520. Singer, K., Chang, P.T., Hsu, G.L., 1972. Physique, personality and mental illness in the Southern Chinese. Br. J. Psychiatry 121, 315–319. Sorensen, H.J., Mortensen, E.L., Reinisch, J.M., Mednick, S.A., 2006. Height, weight and body mass index in early adulthood and risk of schizophrenia. Acta Psychiatr. Scand. 114 (1), 49–54. Sugawara, N., Yasui-Furukori, N., Tsuchimine, S., et al., 2012. Body composition in patients with schizophrenia: comparison with healthy controls. Ann. General Psychiatry 11. https://doi.org/10.1186/1744-859X-11-11. Susser, E., St Clair, D., 2013. Prenatal famine and adult mental illness: interpreting concordant and discordant results from the Dutch and Chinese famines. Soc. Sci. Med. 97, 325–330. Woerner, M.G., Pollack, M., Klein, D.F., 1971. Birth weight and length in schizophrenics personality disorders and their siblings. Br. J. Psychiatry 118 (545), 461–464. Xu, T., Chan, R.C., Compton, M.T., 2011. Minor physical anomalies in patients with schizophrenia, unaffected first-degree relatives, and healthy controls: a meta-analysis. PLoS One 6 (9), e24129.
Please cite this article as: Latham, K., Kirkpatrick, B., Meta-analysis of adult height and birth length in schizophrenia, Schizophr. Res. (2017), https://doi.org/10.1016/j.schres.2017.09.002