The impact of early life factors on cognitive function in old age: The Hordaland Health Study (HUSK)

Study population

The sampling frame for this study comprised the participants of the old age cohort of the population-based Hordaland Health Study (HUSK) which has been described in more detail elsewhere (Refsum et al. 2006). In summary, all residents of Bergen city or neighbouring areas born during the period of 1925–27 of a previously established cohort were invited to participate in a general physical examination and to complete a set of questionnaires on socio-demographic status, general health and health-related behaviour. HUSK was conducted from 1997 to 1999 as a collaboration between the National Health Screening Service, the University of Bergen and the local health services. A random subsample of the attendees in the old age cohort (n=3,341) was also invited to participate in a cognitive examination, with 2,203 (66% of the attendees) agreeing to participate (Figure 1). Of these, 2,156 had complete data and were included in analyses presented here.

In the Norwegian Population Registry, all inhabitants of Norway are registered with a personal identification number. Using this individual identifier, the names (and maiden name for females), date of birth, place of birth and parents’ names (when available) of HUSK participants were retrieved. This information was used to trace the participants born in Bergen to the birth records from the public maternity ward (“Fødestiftelsen i Bergen”) presently stored at the Regional State Archives of Bergen. In the second decade of the 20th century, about one quarter of all births in the Bergen area took place in the official maternity ward (personal communication, State archivist). The proportion of deliveries taking place at hospitals increased steeply when the new Women’s Clinic (“Kvinneklinikken”) was inaugurated in 1926, replacing the old maternity ward. The pertinent birth records for the present study were those detailing births between 1st of January 1925 and 31st of December 1927, and these records have been employed previously in a similar study design (Skogen et al. 2013). The records contain detailed information about the pregnancy, the birth process and the mother’s health recorded by midwives and obstetricians during the hospital stay. The Women’s Clinic in question was the main teaching facility for midwifes at the time, and the records were requisite for the training, and are therefore considered to be of high quality (Rosenberg 1987). Of the 2,156 participants in the HUSK cognitive examination, we were able to trace 346, which constituted the final study sample aged 72–74 years (mean 72.3).

Early life factors – information obtained at birth, 1925–27

The available birth records in the Regional State Archives of Bergen were viewed and coded blind to all HUSK measures. The following information was abstracted from the record (directly copying original information unless stated otherwise): birth weight (kg), birth length (cm), head circumference (cm) at birth, ponderal index (PI; calculated from weight and length), mother’s pelvic size (the mean of the interspinous distance, the intercristal distance and the external conjugate in centimeters). The following binary variables were derived from individual free text fields: any recorded disease in the mother (yes/no), family history of coronary heart disease (yes/no) and tuberculosis (TB; yes/no), the state of mother’s teeth (poor/good), mother’s condition after birth (poor/good), complications during birth (including, but not limitied to, prolonged labour, abnormal presentation, assisted delivery of the baby (use of forceps) and episiotomy, uterine rupture, discoloured amniotic fluid, abnormal fetal souffle and placenta praevia; yes/no), mother’s general somatic state at discharge (poor/good), marital status (married/unmarried), socioeconomic status (based on father’s occupation; lower/higher), and type of payment for the hospital stay (health insurance/other).

Cognitive examination at age 72–74 years of age

HUSK included a cognitive test battery consisting of six tests. The cognitive tests are in wide use internationally and have been well validated, including the Norwegian versions of MMSE and KOLT (Kendrick 1985; Wechsler 1981; Benton & Hamscher 1989; Braekhus et al. 1992; Reitan 1958; Engedal et al. 1988). Two assessors were trained over two days to use the test battery (personal communication, Professor Knut Engedal). These assessors were nurses, and the battery was administered on-site by the trained nurses at the end of the study’s examination.

Context for the birth cohort

During the late 19th century and early 20th century, Bergen city expanded geographically, and went from a semi-rural city to a city with more modern characteristics. Primary industry which had dominated gave way for an expanding secondary and tertiary industry (Ertresvaag 1982). This change in industry was mostly due to growing production and manufacturing, but also due to an increase in commerce, shipping and transport, and service sector (Ertresvaag 1982). As a consequence of this, three social classes began to dominate in Bergen during the same period, upper (bourgeoisie), middle and lower, with large differences in income, housing standard and diet. The upper class was characterised by financers, importers, industry proprietors and wholesale dealers. The middle class consisted primarily of craftsmen, merchants and officials, while the lower class comprised regular worker or artisans (Ertresvaag 1982). During 1925 and 1927 the life expectancy in Norway was approximately 67 years for males, and 74 years for females (Mamelund & Borgan 1996).

Additional information gathered during follow-up from HUSK at age 72–74

Potential differences in the distribution of gender, self-reported level of educational attainment and general health were investigated between the HUSK participants with birth journal information (N=346) and participants without (N=1,810). Level of educational attainment was divided into “compulsory only” (up to ten years) and “post-compulsory” (11 years or more), while general health was divided into “poor” and “good”. As APOE gentotype has been associated with cognitive function (Izaks et al. 2011), information about apoE4-status (presence of any E4-allele versus absence of E4-allele) was also included (using nonfasting plasma samples taken during the general physical examination of HUSK).

Statistical analyses

HUSK participants with traceable birth records were compared to the remainder of the HUSK participants. Bivariate and age- and gender-adjusted associations were then investigated between exposures and outcomes employing linear regression models. Our approach was to investigate and report all associations between exposures and outcomes, taking into account the number of significant associations that would be expected through chance alone, but also evaluating the output for any consistency in associations for a given exposure or outcome (Rothman 1990). For the main analysis, Stata version 11.0 (StataCorp 2010) was employed. Using the software G*Power version 3.1.3 (http://www.psycho.uni-duesseldorf.de/abteilungen/aap/gpower3/) a power analysis indicated that we would be able to detect a small to medium effect size for continuous outcomes (a correlation of 0.13), and mean differences (Cohen’s d of 0.35) at a power of 80% (alpha 0.05) given our sample size (Cohen 1992). We also investigated the potential two-way interaction between apoE4-status and gender for each of the exposures in relation to the composite score, in a post-hoc analysis. Post-hoc analyses were also performed to investigate whether the effect of parental SES on cognitive function were independent of anthropometric measures, as well as whether the effect of head circumference on cognitive function was independent of parental SES. In sensitivity analyses, we also explored the effect of separate additional adjustment for educational attainment and self-rated general health on those associations found to be significant after age- and gender-adjustment.

Ethics

The data in HUSK was collected in accordance with ethical standards required by the regional ethical board of Committees for Medical and Health Research Ethics in Norway (REC). The permission to collect and store the data from HUSK was given by the Norwegian Data Inspectorate. All participation in HUSK was voluntary, and all potential participants received written information about the project before they met for examination. The participants gave their written statements of informed consent, including the specific consents to use information from HUSK in health research and to link this information with other relevant data sources. This specific study was reviewed and approved by REC.

Interpretation of our findings

We found little evidence to support a substantial association between intrauterine or birth environment and cognitive function in old age in general. The only anthropometric measure which to a certain degree predicted cognition in old age was head circumference, and parental SES was the only exposure which was associated with the composite cognitive score. Both a higher head circumference and SES seemed to predict a higher cognitive function in old age independently of each other. One potential explanation for this is that early SES and head circumference are predictors of two different aspects of later cognitive function (Stern 2002). The association between SES and later cognitive function may represent cognitive reserve, while the association between head circumference and later cognitive function may represent brain reserve, both of which are relevant concepts for understanding cognitive function and vulnerability to cognitive impairments in old age (Stern 2002). In this respect, it also interesting that adjusting for educational attainment substantially weakened the associations between paternal SES and cognitive function in old age, suggesting that these associations might be substantially mediated through education. On the other hand, self-rated health reported in later life did not appear to influence these associations meaningfully. Further specific causal pathway modeling was felt to be beyond the scope of this study and not warranted by the largely negative associations of interest.

The lack of a substantial association between intrauterine or birth environment and cognitive function in old age, may be also be a reflection of a diminished impact of these early factors as other influences comes into play across the lifespan (Zhang et al. 2009). Both birth weight and socioeconomic status have been found to be associated with cognitive function in childhood (Shenkin et al. 2001; Jefferis et al. 2002), although socioeconomic status and postnatal influences have been suggested to be more important than prenatal factors (Jefferis et al. 2002; Erickson et al. 2010), similar to our own finding of the importance of parental SES. Other studies have also found that social disadvantage and early life stressors are related to cognitive function in later life (Mak et al. 2006; Nguyen et al. 2008; Fors et al. 2009), and it is generally accepted that childhood SES is an important predictor for later cognitive function (Mak et al. 2006; Hackman & Farah 2009), and cognitive reserve (Stern 2002). Even though anthropometric measures obtained at birth did not predict cognitive function later in life, it is possible that other factors mitigated these initial differences and reduced or eliminated their influence in later adult life (Zhang et al. 2009). These may include later nutrition, education and occupational status (Stern 2002).