A 16-Year Prospective Study of Prodromal Features Prior to BPI Onset in Well Amish Children

Abstract

Background

Longitudinal research of well Amish children over 16 years to identify the pattern and frequency of prodromal symptoms/behaviors associated with onset of BPI disorder during childhood or adolescence. Methods: Parental informants were interviewed annually using structured and semi-structured interviews to record medical, developmental and behavioral/symptomatic data for their children in two samples. The bipolar sample had 115 children with a BPI parent. The control sample had 106 children of well parents, with and without a positive family history for mood disorders. A panel of clinicians assigned risk ratings independently and blind to family relations. Results: Eight children, age 13 or older, onset with BPI in the bipolar sample compared with one in the control sub-sample (well parent of a BPI sibling). The specific “pre-school” behaviors/symptoms that most identified children with BPI from well children in control samples were: sensitivity, crying, hyper-alertness, anxiety/worry and somatic complaints. During school years, parents reported mood (sad) and energy changes (low not high) decreased sleep and fearfulness as key symptoms. Limitations: The sample of 9 BPI onsets is small. However, a variable age of onset means many children remain at risk. Although not statistically significant, 34.6% of the bipolar sample youngsters carry risk ratings compared to 15.2% among controls. Conclusions: The miniclusters of prodromal features that emerged pre-school (ages 1–6), were “episodic” through childhood (7–12) and appeared to mimic adult recurrent illness. Prepubertal onset with mania did not occur. The pattern of prodromal symptoms has clinical relevance for its potential predictive value for onset with BPI disorder and early intervention.

Introduction

The Amish Study, launched in 1976, was described as an ideal “naturalistic laboratory for epidemiologic and genetic research” that could address the natural history of major affective disorders in large informative families with a longitudinal approach (Andreasen (1983)). Even though some clinicians were skeptical of the concept of a “depressed” child during the 1970–1980s, serious research was being done on depressed children and a review summarized the then current thinking about major depression among children (Carlson and Cantwell, 1980, Weller and Weller, 1984).

The historic concept of the “bipolar” child can be found in the case study of a boy who had a first manic episode at age 12, described as a latent “manic-depressive tendency” but not yet documented as a childhood disorder (Anthony and Scott, 1960, Anthony and Scott, 1960). Their extensive literature review yielded several prepubertal cases that met their criteria and helped usher in debate about early childhood symptoms for mania and pubertal maturation. Investigation into the onset and course of illness for referred juveniles and young offspring of adult BPI patients demonstrated high-risk familial loading (Akiskal et al., l985). Furthermore, this research reported that no “full-blown” mania was observed prior to puberty. Results from a study of “age of onset criteria” found first noticeable affective symptoms could be coded 9–12 years before full onset (Egeland et al., 1987). Subsequently, a retrospective study of first hospital admissions for 58 Amish adults with BPI confirmed this time lapse between first symptoms and onset with the full syndrome (Egeland et al., 2000). The Amish Study also reported that age-corrected lifetime rates of illness for 408 first degree relatives of Amish BPI probands were highest among the offspring (Pauls et al., 1992).

A shortcoming of most early studies was the retrospective nature of the work. While the need for prospective research was clear, the cost of conducting well designed longitudinal studies made it difficult to obtain funding. Results from the few prospective studies (late 1980s–early 1990's) confirmed the increased prevalence of pediatric bipolarity among children of BPI parents (Zahn-Waxler et al., l988). In a more recent study, Duffy and colleagues reported findings from a 15-year longitudinal study of children at risk for bipolar disorder (Duffy et al., 2009). Results replicated the increased risk of BPI but noted low rates for attention-deficit/hyperactivity disorder (ADHD), anger and disruptive behaviors, higher rates for anxiety and sleep problems, and absence of prepubertal mania. Many of these findings are consistent with earlier reports from the Amish Study (Egeland et al., 2003, Shaw et al., 2005) and with the 16-year longitudinal data reported in this paper.

The prospective Children and Adolescent Research Evaluation (CARE) study was initiated in 1994. It had full cooperation of Amish parents concerned about early detection of bipolar symptoms. They understood manic-depressive disorder to be an inherited medical problem prevalent in certain extended families. The “at risk” children in the CARE study represent the 4^th –5^th generations in pedigrees with high loading for classic BPI disorder. This genetic context was central to our “a priori” hypothesis. The families were recruited directly from existing linkage pedigrees that have been under intensive study, employing “state of-the-art” molecular genetic strategies available in the late 20^th century (Ginns et al., 1996). In the 7^th and 10^th Year CARE articles, significant differences were reported for risk and possible antecedents between well children in the bipolar sample and well children in a control sample (Egeland et al., 2003, Shaw et al., 2005). There was an interesting shift from internalizing symptoms to more externalizing ones as children reached puberty. Miniclusters of possible early bipolar predictors were said to “go underground” and emerge periodically. This episodic pattern was in contrast to the chronic, continuous course of illness, complicated by co-morbidity, previously reported by other studies. Full-blown manic symptoms did not appear until after puberty. During the time of the design, support and execution of the CARE program, the pendulum has swung from a disbelief in childhood bipolarity to “variations” in almost Paganini proportions. This paper is the 16^th Year report from a “naturalistic laboratory” addressing risk and onset of BPI among Amish children who were well at the initiation of the study and followed longitudinally from 1994 through 2009.