Introduction
One of the most important developments in translational mental health is the observation that the inflammatory system is involved in the pathogenesis of major depressive disorder (MDD) (Bufalino et al. Reference Bufalino, Hepgul, Aguglia and Pariante2013; Valkanova et al. Reference Valkanova, Ebmeier and Allan2013; Zunszain et al. Reference Zunszain, Hepgul and Pariante2013; Kiecolt-Glaser et al. Reference Kiecolt-Glaser, Derry and Fagundes2015; Miller & Raison, Reference Miller and Raison2015). Activation of the immune system in subsets of depressed patients plays a role not only in disease progression, but also in determining the success of antidepressant therapy (Zunszain et al. Reference Zunszain, Anacker, Cattaneo, Carvalho and Parainte2011, Reference Zunszain, Hepgul and Pariante2013; Haroon et al. Reference Haroon, Raison and Miller2012; Strawbridge et al. Reference Strawbridge, Arnone, Danese, Papadopoulos, Herane Vives and Cleare2015). However, the temporal relationship remains largely unclear: is increased inflammation a cause or an effect of MDD?
Interestingly, exposure to increased inflammation may indeed play a causal role in the pathogenesis of depression, as shown from research detailing how a high proportion of patients undergoing treatment with interferon (IFN)-α, a cytokine used for cancer or viral hepatitis C, go on to develop depression (Capuron et al. Reference Capuron, Ravaud, Gualde, Bosmans, Dantzer, Maes and Neveu2001; Bonaccorso et al. Reference Bonaccorso, Marino, Biondi, Grimaldi, Ippoliti and Maes2002; Horikawa et al. Reference Horikawa, Yamazaki, Izumi and Uchihara2003; Loftis & Hauser, Reference Loftis and Hauser2004; Raison et al. Reference Raison, Demetrashvili, Capuron and Miller2005). Furthermore, various longitudinal studies support that diabetes mellitus, obesity, and cardiovascular disease, all characterized as low-grade chronic inflammatory states, are significant predictors of depression in later life (Mezuk et al. Reference Mezuk, Eaton, Albrecht and Golden2008; Nabi et al. Reference Nabi, Kivimäki, Suominen, Koskenvuo, Singh-Manoux and Vahtera2010; Hare et al. Reference Hare, Toukhsati, Johansson and Jaarsma2014; Luppino et al. Reference Luppino, De Wit, Bouvy, Stijnen, Cuijpers, Penninx and Zitman2015).
However, not everyone exposed to increased inflammation develops depression. As with all other diseases, there are risk factors that contribute to an increased vulnerability in certain individuals. Such risk factors could include the type, severity, frequency and/or the timing of an inflammatory challenge. Indeed, research has already shown that the timing and/or age at exposure may be an important predictor of future psychopathology, with exposure to adverse experiences particularly in early life being consistently associated with increased susceptibility to a variety of neuropsychiatric disorders (Turecki et al. Reference Turecki, Ota, Belangero, Jackowski and Kaufman2014; Visser et al. Reference Visser, van Minnen, van Megen, Eikelenboom, Hoogendoorn, Kaarsemaker, Balkom and van Oppen2014; Kalmakis & Chandler, Reference Kalmakis and Chandler2015; Trotta et al. Reference Trotta, Murray and Fisher2015). In two prospective studies, although depression in adulthood was associated with an accumulation of stressors across the life course, most originated in the first years of life (Clark et al. Reference Clark, Caldwell, Power and Stansfeld2010; Colman et al. Reference Colman, Jones, Kuh, Weeks, Naicker, Richards and Croudace2014). Furthermore, childhood adversity was either directly associated with adolescent, early adulthood, and mid-life affective disorder psychopathology (Clark et al. Reference Clark, Caldwell, Power and Stansfeld2010), or was associated with intermediate risk factors that subsequently increased the risk of future depression (Colman et al. Reference Colman, Jones, Kuh, Weeks, Naicker, Richards and Croudace2014). Thus, it would seem that adversity in early life has important effects on the life course of depression, and that timing of such adversity may be an important factor in the aetiology of the disorder.
Admittedly, there may even be a time point in early life when individuals are most vulnerable to adverse exposure. In 2012, Bosch and colleagues demonstrated how the sequelae of early life adversity depended on the age at the time of exposure, showing how timing of an adverse event could differentially alter the functioning of the hypothalamic-pituitary-adrenal (HPA) axis in later life. Specifically, they showed how hypercortisolism was a potential consequence of adverse exposure between 6 and 11 years, while adverse exposure between 12 and 15 years contributed to hypocortisolism. Interestingly, individuals exposed before the age of five had no alterations in stress responsivity (Bosch et al. Reference Bosch, Riese, Reijneveld, Bakker, Verhulst, Ormel and Oldehinkel2012). This suggests that the first 5 years of life could represent a stress-hypo-responsive period (Sapolsky & Meaney, Reference Sapolsky and Meaney1986) – a temporary and well-documented developmental period in life characterized by attenuated stress responsivity (Gunnar et al. Reference Gunnar, Brodersen, Krueger and Rigatuso1996; Larson et al. Reference Larson, White, Cochran, Donzella and Gunnar1998; Gunnar & Donzella, Reference Gunnar and Donzella2002) – the evolutionary purpose of which is thought to promote maternal-infant attachment, with low levels of cortisol attributed to maintaining and reinforcing this attachment (Moriceau & Sullivan, Reference Moriceau and Sullivan2007). Poignantly, impaired maternal-infant attachment has been associated with poor emotional outcomes in offspring (Leckman-Westin et al. Reference Leckman-Westin, Cohen and Stueve2009), so further understanding of this physiological phenomenon could help extend our knowledge on the aetiology of depression. Indeed, the idea that the stress system (and perhaps even the immune system) may be biologically more or less responsive depending on developmental age could explain why certain types of exposures at certain times in life may exert differential outcomes, and this study emphasizes not only the importance of the type of adversity and the frequency of its occurrence, but also the timing of exposure.
To date, no clinical study has established whether there exists a critical period in life when adversity in the form of a medically-related immune insult may increase one's susceptibility to mental illness. However, one of the hallmarks of the developing immune system is that it exhibits an increased sensitivity for environmentally induced toxicity compared to the fully matured immune system of an adult (Dietert et al. Reference Dietert, Etzel, Chen, Halonen, Holladay, Jarabek, Peden, Pinkerton, Smialowicz and Zoetis2000; Holladay & Smialowicz, Reference Holladay and Smialowicz2000; Dietert, Reference Dietert2013). Moreover, an early life inflammatory insult can result in the impairment of a variety of biological systems involved in the aetiology of MDD, including the neuroendocrine system (Rivest, Reference Rivest2010). Indeed, it seems highly plausible that a ‘critical window’ of vulnerability exists for immune system activation on mental health susceptibility, and that this window of vulnerability could exist when the immune system is still maturing in early life, which starts in utero, and continues until the age of 15 years (Hannet et al. Reference Hannet, Erkeller-Yuksel, Lydyard and Debruy1992; Osugi et al. Reference Osugi, Hara, Kurahashi, Sakata, Inoue, Yumura-Yagi, Kawa-Ha, Okada and Tawa1995; Holt & Jones, Reference Holt and Jones2000).
Review objectives
The primary objective of this article is to review the current clinical literature in order to elucidate whether exposure to an inflammatory insult early in life, driven by medical or infective causes, increases the risk for depression later in life. Given that neural development extends from the embryonic period through to adolescence (Rice & Barone, Reference Rice and Barone2000; Johnson, Reference Johnson2001), and that this coincides with the development of the immune system (Hannet et al. Reference Hannet, Erkeller-Yuksel, Lydyard and Debruy1992; Osugi et al. Reference Osugi, Hara, Kurahashi, Sakata, Inoue, Yumura-Yagi, Kawa-Ha, Okada and Tawa1995; Holt & Jones, Reference Holt and Jones2000) we focus on studies reporting exposure to increased inflammation in three developmental life stages: antenatal, childhood (birth to age 12), and adolescence (age 13–18). We specifically exclude studies where inflammation was driven by exposure to psychosocial stress or adversity. Being able to identify when during the developmental trajectory an immune insult is more detrimental could have significant implications in terms of developing successful prevention strategies, raising awareness, and targeting more vulnerable individuals, and as such, it is important to establish whether timing does indeed matter.
Method
Search strategy and limits
A combination of PubMed, EMBASE, Ovid Medline and PsycINFO databases were used to systematically select studies for discussion, and reference lists of selected papers were manually searched to check for any additional studies. An independent systematic search of all aforementioned databases was carried for each life stage under study.
We included only clinical studies that were longitudinal in nature, and where there was a minimum of 2 years between exposure and mental health assessment, thereby minimizing any potential contamination of measures. No publication date restrictions were imposed, but our searches were limited to English-language studies only.
Inflammatory exposure: definition and limits
We defined an inflammatory challenge as any illness pertaining to increased inflammation, which we classified as either a direct, or indirect, immune challenge.
A direct immune challenge was characterized as any bacterial, fungal, parasitic, allergenic and/or viral infection/illness, either chronic or acute, in which the primary host's response to infection/illness was inflammation. Chronic illnesses resulting in the sustained use of immunosuppressant medication, or any autoimmune disease/condition resulting in organ transplantation were excluded.
An indirect immune challenge was defined as any exposure to an illness/condition characterized as a systemic inflammatory state. Substantial evidence chronicles the activation of the immune system in diabetes mellitus (gestational, types I and II) (Donath & Shoelson, Reference Donath and Shoelson2011; Calle & Fernandez, Reference Calle and Fernandez2012), obesity (Kredel & Siegmund, Reference Kredel and Siegmund2014; de Jong et al. Reference de Jong, Kloppenburg, Toes and Ioan-Facsinay2014), congenital heart disease (Sharma et al. Reference Sharma, Bolger, Li, Davlouros, Volk, Poole-Wilson, Coats, Gatzoulis and Anker2003; Allan et al. Reference Allan, Newburger, McGarth, Elder, Psoinos, Laussen, del Nido, Wypij and McGowen2010), and cardiovascular disease (Hansson, Reference Hansson2005; Mangge, Reference Mangge2014). Reciprocity between inflammation and these conditions have been consistently demonstrated, such that they are typically referred to as chronic low-grade inflammatory states. As such, we included all studies reporting (a) exposure to maternal obesity and diabetes antenatally (in utero), and (b) living with diabetes, cardiovascular disease, and/or obesity postnatally, in relation to the development of depression in later life. Although cancer is a well-known inflammatory state (Payne, Reference Payne2014; Roxburgh & McMillan, Reference Roxburgh and McMillan2014), due to the vast heterogeneity of this disease, studies on cancer were excluded.
Depressive psychopathology: definition and limits
Depressive psychopathology was conceptualized by the use of affective/mood symptoms and diagnoses. Most studies assessed depression and/or depressive symptomology in adult participants, i.e. aged >18 years, but we also included papers reporting symptoms or diagnoses of depression in adolescent participants, as the effect(s) of early-life adversity on mental health may have emerged by this time point (Pawlby et al. Reference Pawlby, Hay, Sharp, Cerith and Pariante2011, Plant et al. Reference Plant, Pariante, Sharp and Pawlby2015). For studies assessing mental health in adolescence, we also accepted papers reporting symptoms of internalizing and externalizing behaviours, which have been shown to predict adulthood affective disorders (Roza et al. Reference Roza, Sc, Hofstra, Van Der Ende and Verhulst2003; Clark et al. Reference Clark, Caldwell, Power and Stansfeld2010).
For a full list of key words used in our searches see Supplementary online Appendix.
Results
Fig. 1 highlights the number of studies identified at each stage of our search strategy for all stages combined. Of the 10 087 papers initially flagged up, only 22 clinical studies were eligible and ultimately included in our review.
Fig. 1. Flow diagram depicting the search strategy employed, the number of studies identified at each stage, the number of studies excluded at each stage and those included in the review.
Exposure to increased inflammation antenatally
Table 1 displays all clinical studies that directly and indirectly investigated the effect of an antenatal inflammatory challenge on depression susceptibility in later life.
Table 1. Studies examining the association between in utero exposure to inflammation and risk for depression in later life
a No differences in risk of depression.
b Significantly increased risk of depression.
Exposure to maternal infection in utero
Although several studies have associated maternal infections with adult psychiatric disorders, predominantly in schizophrenia, autism spectrum disorder, epilepsy and cerebral palsy (Knuesel et al. Reference Knuesel, Chicha, Britschgi, Schobel, Bodmer, Hellings, Toovey and Prinssen2014), the impact of maternal infection on depression in offspring is less well-known. To date, only six clinical papers examined whether exposure to infection during pregnancy increased risk of offspring depression (Brown et al. Reference Brown, Susser, Lin and Gorman1995; Machón et al. Reference Machón, Mednick and Huttunen1997; Mino et al. Reference Mino, Oshima and Okagami2000; Mellins et al. Reference Mellins, Smith, O'Driscoll, Magder, Brouwers, Chase, Blasini, Hittleman, Llorente and Matzen2003; Gaughan et al. Reference Gaughan, Hughes, Oleske, Malee, Carol, Nachman, Malee, Gore and Nachman2004; Pang et al. Reference Pang, Syed, Fine and Jones2009).
In one such study, a significant increase in depression was reported for individuals exposed during their second trimester to a Finnish influenza epidemic compared to control subjects born 6 years prior to the outbreak (13% v. 2%) (Machón et al. Reference Machón, Mednick and Huttunen1997). In contrast, two studies examining patients with mood disorders born during an influenza epidemic in Holland and Japan found no significant differences in the risk for depression in patients born during the epidemics (Brown et al. Reference Brown, Susser, Lin and Gorman1995; Mino et al. Reference Mino, Oshima and Okagami2000). However, the contradictory findings between these studies could partially be explained by other environmental factors such as diet, and indeed, low incidence rates of inflammatory disease in Greenland Inuit and Japanese people, thought to be attributed to the large consumption of fish containing omega-3 fatty acids, has previously been reported (Simopoulos, Reference Simopoulos2002). Thus, it seems plausible that women in the Japanese cohort may have had diets rich in omega-3 fatty acids, which consequently may have bestowed some protection against depression (Lin & Su, Reference Lin and Su2007; Sublette et al. Reference Sublette, Ellis, Geant and Mann2011; Martins et al. Reference Martins, Bentsen and Puri2012).
Recently, however, a larger UK study examining the effect of a variety of prenatal viral infections and offspring depression found no overall increased risk for depression associated with viral exposure (Pang et al. Reference Pang, Syed, Fine and Jones2009). Similarly, two large studies investigating the impact of antenatal exposure to human immunodeficiency virus (HIV) on the subsequent development of depression in later life found that HIV exposure did not predict the development of poor emotional and/or behavioural outcomes. Although a high prevalence of behavioural problems did exist among HIV-infected children, these studies found that neither HIV infection nor prenatal drug exposure was the underlying cause (Mellins et al. Reference Mellins, Smith, O'Driscoll, Magder, Brouwers, Chase, Blasini, Hittleman, Llorente and Matzen2003; Gaughan et al. Reference Gaughan, Hughes, Oleske, Malee, Carol, Nachman, Malee, Gore and Nachman2004).
Exposure to gestational diabetes and obesity in utero
Interestingly, no studies on the effect of gestational diabetes or obesity on increased risk of offspring depression were found. Although gestational diabetes has been linked to altered brain development and behaviour in offspring, specifically in autism spectrum disorder (Xu et al. Reference Xu, Jing, Bowers, Liu and Bao2014) and attention-deficit hyperactivity disorder (Nomura et al. Reference Nomura, Marks, Grossman, Yoon, Loudon, Stone and Halperin2012), no study was found in relation to depression.
Exposure to increased inflammation in childhood
Clinical studies examining whether direct or indirect exposure to inflammation in childhood predicts depression in later life are given in Table 2.
Table 2. Studies examining the association between childhood exposure to inflammation and risk for depression in later life
a No differences in risk of depression.
b Significantly increased risk of depression.
c Significantly increased risk of depression for particular subset of cohort.
Chronic illness in childhood and depression in later life
Eight studies pertaining to chronic or acute illness in childhood were found (Pless et al. Reference Pless, Cripps, Davies and Wadsworth1989; Kokkonen & Kokkonen, Reference Kokkonen and Kokkonen1993; Cohen et al. Reference Cohen, Pine, Must, Kasen and Brook1998; Packham et al., Reference Packham, Hall and Pimm2002; Gaughan et al. Reference Gaughan, Hughes, Oleske, Malee, Carol, Nachman, Malee, Gore and Nachman2004; Goodwin, Reference Goodwin2011; Ferro & Boyle, Reference Ferro and Boyle2015; Khandaker et al. Reference Khandaker, Pearson, Zammit, Lewis and Jones2014). Of these, five used a broad definition of infection/illness, and as such had heterogeneous diagnostic groups, while the other two referred to a specific infection/illness.
Exposure to heterogeneously defined physical illness in childhood
One British birth cohort investigated the effect of chronic illness in childhood on the mental health wellbeing of participants at ages 26 and 36 years (Pless et al. Reference Pless, Cripps, Davies and Wadsworth1989). The authors found no overall significantly increased reports of psychiatric disorders in chronically ill children compared to healthy controls (men: 6.1% v. 3.8%; women: 14.1% v. 9%). However, they did find that participants who were ill in childhood and again after 21 years of age were significantly more likely to have a psychiatric disorder compared to other members of the cohort. Additionally, when cohort members were re-interviewed at 36 years of age, specifically to assess current affective state, women who experienced childhood chronic illness had significantly higher depressive scores compared to controls. This study highlights that (a) repeated exposure to medically-related inflammation, across two life stages, may be necessary for eliciting psychopathology in some individuals, and (b) the effect of early life illness may operate only for a subset of individuals, with women, in this instance, being more vulnerable.
In a second study, the prevalence rates of mental disorders in adults who suffered from a variety of chronic physical illnesses in childhood were compared to age-matched controls (Kokkonen & Kokkonen, Reference Kokkonen and Kokkonen1993). Similar to findings from Pless and colleagues, this study found no significant difference in the prevalence rates of all types of depression in young adults with childhood illness compared to healthy controls (13% v. 12%). However, severe depression was significantly more common in patients exposed to childhood illness than in controls (6% v. 2%). Again, we find only a subset of vulnerable individuals, and, as in this case, unless studies examine diagnosis by severity, participants with milder forms of depression may mask the association between early life illness and more severe forms of depression.
Contrary to the findings of the two aforementioned studies, which all demonstrate an increased vulnerability to depression in only a subset of participants suffering from illness as children, three more recent studies find an overall significant association between childhood infection and the mental health of their cohort (Cohen et al. Reference Cohen, Pine, Must, Kasen and Brook1998; Goodwin, Reference Goodwin2011; Ferro & Boyle, Reference Ferro and Boyle2015). A large US cohort found that chronic physical illness in childhood predicted an increased risk of future depression in both adolescence [odds ratio (OR) 3.81, 95% confidence interval (CI) 1.55–9.39] and young adulthood (OR 4.04, 95% CI 1.54–10.62) independent of prior depressive episodes and other demographic covariates. Moreover, the authors also showed how immunologically mediated disorders, specifically atopic illness, hay fever and mononucleosis, exhibited strong associations with subsequent onset of depression in both adolescence and young adulthood (Cohen et al. Reference Cohen, Pine, Must, Kasen and Brook1998). Similarly, another large, but retrospective study, examined the association between infection in the first year of life and mental disorders among youth in a community sample, and reported that early life infection was associated with significantly increased odds of depression (OR 3.7, 95% CI 1.0–13.4) (Goodwin, Reference Goodwin2011). However, despite the claim of significantly increased odds of depression, we should be mindful that the confidence intervals included one. Finally, in congruence with both Cohen and colleagues, and Goodwin, another large study found that those chronically ill in childhood reported significantly more symptoms of depression in adolescence compared to healthy controls (OR 2.71 v. 2.36) (Ferro & Boyle, Reference Ferro and Boyle2015).
Exposure to specific physical illness in childhood: taking inflammatory markers into account
Thus far there appears to be fairly strong evidence to suggest that exposure to illness in childhood increases the risk for depression in later life, particularly for a subset of individuals. However, none of the aforementioned studies have directly measured levels of inflammation, a concept we should be mindful of in the context of this review. However, there exist three clinical studies that provide some insight into the direct effect of inflammation on depression susceptibility.
One such study, assessing adults with juvenile idiopathic arthritis (JIA), found that patients with systemic-onset JIA had significantly higher levels of depression (10.7%) compared to other JIA subsets. Interestingly, the study also found that depression was most commonly seen when the age of onset was between 6 and 12 years (11.1%) compared to early (2.7%) or late (0%) onset, and that the first episode of depression tended to be between ages 15 and 25 years (38.5%). However, although these findings are highly novel in that they are the first to demonstrate that JIA onset specifically in childhood may exert a greater influence over the subsequent development of depression, the study additionally highlighted how joint inflammation based on the Thompson–Kirwan scale, likely representing the magnitude of systemic inflammation, was not a significant predictor of depression in this cohort (Packham et al., Reference Packham, Hall and Pimm2002). Similarly, another study examining the long-term effects of postnatal exposure to HIV on mental health in later life found that although HIV infected children were at increased risk for depression compared to healthy controls, immunological and virological markers were not responsible for predicting first admission hospitalizations for depression (Gaughan et al. Reference Gaughan, Hughes, Oleske, Malee, Carol, Nachman, Malee, Gore and Nachman2004). Interestingly, these studies emphasize how psychological variables, rather than the acute effects of specific inflammatory factors, may explain the majority of variance seen in later life depression.
However, one population-based study showed how participants with increased levels of systemic inflammation in childhood, indicated by higher serum levels of interleukin-6, were at significantly increased risk of developing depression in adulthood (Khandaker et al. Reference Khandaker, Pearson, Zammit, Lewis and Jones2014). Unlike many of the already discussed papers, this study's strength was that it controlled for a wide variety of confounders including past psychological and behavioural problems and maternal psychopathology, potentially controlling for adversity-related causes of inflammation, as well as children with medical conditions. Therefore, these findings are unique in that they show how raised levels of inflammation, not likely accounted for by adversity, infection or illness, can predict future psychopathology. This suggests that perhaps an inherently dysfunctional immune system rather than exposure to inflammation via infection per se is the key to understanding the association between inflammation and depression.
Living with obesity, diabetes mellitus and congenital heart disease in childhood and later life depression
Widening our search to incorporate studies investigating the long-term impact of childhood obesity, congenital heart disease, and type I diabetes mellitus yielded an additional three papers.
Childhood obesity and depression in later life
Although previous studies have posited a link between adiposity and depression, with inflammation playing a key role in the disorder's pathogenesis (Shelton & Miller, Reference Shelton and Miller2011), no published clinical studies investigating the impact of childhood obesity on future depression were found. However, this is unsurprising since only recently has the impact of obesity in relation to mental health been under full investigation. Given the time and resources involved in conducting longitudinal studies, insights into the impact of childhood obesity on mental health in later life is likely to emerge in subsequent years.
Type I diabetes mellitus and depression in later life
Two clinical studies investigating the effect of diabetes in childhood on affective disorder psychopathology were found, with contradictory conclusions as to when the sensitive period for disease onset may lie. One study reported an increased risk for depressive disorders in early adulthood for childhood-onset diabetic patients (Kovacs et al. Reference Kovacs, Goldston, Obrosky, Scott and Bonar1997), while the other found that adulthood onset type I diabetic women reported higher levels of depression than childhood-onset diabetic patients (Lašaitė et al. Reference Lašaitė, Ostrauskas, Žalinkevičius and Jurgevičienė2015).
Congenital heart disease and depression in later life
The influence of congenital heart disease on later life affective disorder psychopathology has been investigated in only one longitudinal study, which found no overall difference in the mental health outcomes of these patients and healthy controls. However, similar to previous findings from other studies, the study did find that for a subset of the cohort – female patients, and those with more complex forms of the disorder – significantly higher levels of anxiety and depression were reported. Moreover, the authors found that age at assessment was important for evaluating the impact of these disorders on later mental health, finding that those aged 19–26 years had more symptoms of anxiety/depression than those aged 12–18 years (Areias et al. Reference Areias, Pinto, Vieira, Teixeira, Coelho, Freitas, Matos, Castro, Sarmento, Viana, Quintas and Areias2013).
Exposure to increased inflammation in adolescence
Table 3 displays all clinical studies that have assessed whether direct or indirect exposure to increased inflammation in adolescence predicts future depression.
Table 3. Studies examining the association between adolescent exposure to inflammation and risk for depression in later life
a No differences in risk of depression.
b Significantly increased risk of depression.
c Significantly increased risk of depression for particular subset of cohort.
Chronic illness in adolescence and depression in later life
A comprehensive search of the literature yielded one study exploring a direct link between adolescent infection/illness and the development of future depression. However, whilst the data for infection during adolescence was lacking, we found one large study evaluating the relationship of allergic rhinitis (AR) to the development of any depressive disorder in later life (Chen et al. Reference Chen, Su, Chen, Hsu, Huang, Chang and Bai2013). In this study, adolescents with AR had a significantly higher prevalence of major depression (2.5% v. 1.2%) and any depressive disorder (4.9% v. 2.8%) in later life compared to control subjects.
Living with obesity, diabetes and cardiovascular disease in adolescence and depression in later life
Thus far only one study pertaining to adolescent direct exposure has been identified, which has limited our ability to identify any patterns. Broadening our search to incorporate studies looking at the effect of indirect inflammatory conditions yielded an additional five papers.
Adolescent obesity and depression in later life
Three studies exploring the relationship between obesity in adolescence and the subsequent development of future depression were identified. One study reported an overall positive association between obesity and depressive symptoms in adulthood, finding that a higher body mass index (BMI) at age 14 correlated with higher BMI, leptin, C-reactive protein, and depressive symptoms at age 17. Moreover, the study found that females who were obese in both adolescence and adulthood more frequently reported symptoms of depression (Herva et al. Reference Herva, Laitinen, Miettunen, Veijola, Karvonen, Läksy and Joukamaa2006). Interestingly, the remaining two studies found no overall association between adolescent obesity and depression, but did find that the increased risk for the development of future depression was gender specific. Anderson and colleagues found that adolescent obesity in females, but not in males, predicted an increased risk for the subsequent development of depression and anxiety disorder (Anderson et al. Reference Anderson, Cohen, Naumova, Jacques and Must2007). Similarly, Marmorstein and colleagues reported how only obesity in female adolescents predicted the onset of depression in early adulthood: specifically, it was an onset of obesity after the age of 14 that predicted the development of depression in early adulthood among females in the cohort (Marmorstein et al. Reference Marmorstein, Iacono and Legrand2014).
Type I diabetes mellitus and depression in later life
With respect to the impact of diabetes on depression susceptibility in later life only two clinical studies were identified. Both studies found no difference in the psychological outcome of adults diagnosed with diabetes in adolescence compared to healthy controls (Jacobson et al. Reference Jacobson, Hauser, Willett, Wolfsdorf, Dvorak, Herman and de Groot1997; Lašaitė et al. Reference Lašaitė, Ostrauskas, Žalinkevičius and Jurgevičienė2015). Interestingly, however, Jacobson and colleagues did find that in early adulthood individuals with diabetes had lower self-esteem (Jacobson et al. Reference Jacobson, Hauser, Willett, Wolfsdorf, Dvorak, Herman and de Groot1997), a considered predisposing factor for depression (Ferro & Boyle, Reference Ferro and Boyle2015). Therefore, it is possible that these individuals may develop depression in later adulthood.
Cardiovascular disease in adolescence and depression in later life
Regarding the effect of cardiovascular disease in adolescence on mental health, no studies were found. However, this was anticipated given that the onset of cardiovascular disease is typically in adulthood.
Conclusion
Main findings
We have reviewed all available literature examining the effect of a medically related inflammatory challenge in early life, i.e. antenatally, in childhood and in adolescence, on depression susceptibility in later life. We found no clinical evidence to support that inflammation in utero contributes to an increased risk of developing future depression. This was somewhat surprising given that (a) animal research consistently supports that increased inflammation neonatally increases depressive-like behaviour in later life (Walker et al. Reference Walker, March and Hodgson2004, Reference Walker, Hodyl, Krivanek and Hodgson2006, Reference Walker, Knott and Hodgson2008, Reference Walker, Nakamura, Byrne, Naicker, Tynan, Hunter and Hodgson2009; Bilbo et al. Reference Bilbo, Levkoff, Mahoney, Watkins, Rudy and Maier2005; Spencer et al. Reference Spencer, Martin, Mouihate and Pittman2006, Reference Spencer, Galic and Pittman2011; Galic et al. Reference Galic, Riazi, Heida, Mouihate, Fournier, Spencer, Kalynchuk, Teskey and Pittman2008; Roque et al. Reference Roque, Mesquita, Palha, Sousa and Correia-Neves2014), and (b) stress, particularly during pregnancy, is a potential predictor for later life psychopathology (Betts et al. Reference Betts, Williams, Najman and Alati2015; Slykerman et al. Reference Slykerman, Thompson, Waldie, Murphy, Wall and Mitchell2015; Biaggi et al. Reference Biaggi, Conroy, Pawlby and Pariante2016), which is pertinent given the bi-directional relationship between stress and inflammation (Chovatiya & Medzhitov, Reference Chovatiya and Medzhitov2014; Slavich & Irwin, Reference Slavich and Irwin2014). However, given that cytokines do not cross the placenta barrier in normal term foetuses, this may contribute for our findings (Zaretsky et al. Reference Zaretsky, Alexander, Byrd and Bawdon2004; Aaltonen et al. Reference Aaltonen, Heikkinen, Hakala, Laine and Alanen2005).
However, we did find some converging evidence to support that exposure to increased inflammation in childhood increases the risk for adult depression. Furthermore, the evidence suggests that persistent physical health problems in childhood may relate to the presence of greater or more severe psychiatric disorders, and that psychiatric outcome may be gender specific, with females being more vulnerable to exposure. Interestingly, the evidence pertaining to studies assessing exposure to severe conditions in childhood was mixed, but did seem to suggest that for lifelong conditions, such as HIV, diabetes and congenital heart diseases, increased inflammation does not increase susceptibility to depression. However, it is important to note that, for such conditions, treatment strategies may be reducing the overall increased inflammation associated with disease state. Finally, when looking at the direct effect of inflammation in childhood on the susceptibility to depression, the evidence was both limited in quantity and inconsistent in findings, emphasizing the need for further clinical research directly measuring inflammatory markers and evaluating their role in the aetiology of depression.
Looking at adolescent exposure to increased inflammation on future depression risk, a limited quantity of evidence was found. However, we did find some evidence to support that an indirect immune challenge in the form of obesity may increase the risk for future depression, particularly for female adolescents.
Limitations of the existing literature
It is noteworthy that fewer clinical studies were identified for the antenatal and adolescent life stages, and this made it extremely difficult to establish consistent patterns pertaining to these developmental phases from the available evidence. Although we did find some evidence to support that increased inflammation in childhood, and to some extent in adolescence, increases depression risk in later life, many of these studies did not look specifically at levels of inflammation in their cohorts, and as such we cannot conclude that it was increased inflammation per se that predicted depression. Indeed, inflammatory markers were unavailable for the majority of the studies reviewed, predominately because their research question was not addressing the effect of increased inflammation on depression, and as such, this represents a major empirical limitation of conclusions in this review. Other psychological or biological factors associated with these disease states may be accounting for the observed associations.
Furthermore, there are several other limitations that weaken the evidence overall. First, most studies relied on self-reports and/or parental reports of physical illness and affective disorder psychopathology, and it is uncertain whether similar findings would be found with physician-diagnosed medical illnesses. Second, in some studies neither the definition nor timing of physical illness or psychopathology was clearly described. Indeed, several studies assessed exposure to infection across multiple developmental stages, i.e. exposure any time from childhood to early adulthood, and given the large biological and psychosocial differences between the developmental stages, results should be interpreted with caution. Third, important confounders, such as perinatal complications, parental mental health, stress (biological and psychosocial), and childhood maltreatment were not measured and controlled for in analyses in most of the studies. We should be particularly cautious in attributing the observed associations found, especially given the evidence substantiating the involvement of the HPA axis (Pariante & Lightman, Reference Pariante and Lightman2008), childhood maltreatment (Pawlby et al. Reference Pawlby, Hay, Sharp, Cerith and Pariante2011; Lindert et al. Reference Lindert, Von Ehrenstein, Grashow, Gal, Braehler and Weisskopf2014; Plant et al. Reference Plant, Pariante, Sharp and Pawlby2015), parental psychopathology (McLaughlin et al. Reference McLaughlin, Gadermann, Hwang, Sampson, Al-Hamzawi, Andrade, Angermeyer, Benjet, Bromet, Bruffaerts, Caldas-de-Almeida, De Girolamo, De Graaf, Florescu, Gureje, Haro, Hinkov, Horiguchi, Hu, Karam, Kovess-Masfety, Lee, Murphy, Nizamie, Posada-Villa, Williams and Kessler2012), and obstetric complications (Räikkönen et al. Reference Räikkönen, Pesonen, Heinonen, Kajantie, Hovi, Järvenpää, Eriksson and Andersson2008; Tuovinen et al. Reference Tuovinen, Räikkönen, Kajantie, Pesonen, Heinonen, Osmond, Barker and Eriksson2010) in the pathogenesis of depression. Fourth, none of the included studies controlled for infections/illnesses across the life course, and it is possible that the increased risk of depression, found in several studies, may be a consequence of an inflammatory insult at another, more vulnerable, later life-stage, or due to the accumulation of inflammatory insults throughout the life course. Therefore, it is difficult to determine whether the immune system may be ‘primed’ to give an enhanced response after repeated inflammatory episodes throughout life. Indeed, elucidating whether an early life infection can ‘prime’ the developing organism's sensitivity to subsequent environmental challenges is one research question that requires investigation in a clinical setting. Finally, several studies lacked an appropriate control group, making drawing firm conclusions tentative, and/or had a small sample size, potentially lacking the required power to thoroughly investigate the research question.
Despite these limitations, there was considerable agreement with regard to many of the findings pertaining to childhood exposure, and consistency was maintained for studies with both large and small sample sizes, and across individuals with different conditions. Moreover, the lack of an overall association between the prevalence rates of depression in the previously ill v. healthy participants in some of the other aforementioned studies may have been masked by the fact that participants were no older than 25 years at the time of assessment. Empirical research suggests that the average age of onset for mood disorders is 30 years (Kessler et al. Reference Kessler, Berglund, Demler, Jin, Merikangas and Walters2005), and these studies could have potentially failed to capture the expression of depression in such young cohorts. Furthermore, for all prospective studies, there is a degree of confounding by severity, insofar as the participants with the most severe physical and/or mental disorders may not have been able to meet the demands of the studies in question, and a truly representative outcome may not have been achieved.
Future work and implications
Investigating the effect of medically-related inflammation at different life stages will ultimately help identify whether the timing of an immune response is relevant to the pathogenesis of depression, and more prospective longitudinal studies that measure depressive outcomes against number and severity of immune activation throughout life is necessary to confirm this link. Inflammatory markers must be measured in order to investigate the association between severity of immune response and the risk of depression developing. Moreover, controlling for important confounders pertaining to stress (e.g. measuring cortisol stress response), as well as parental psychopathology, and adversity throughout the life course is necessary to confidently establish whether there is a particular time point in life where susceptibility to inflammation predisposes individuals to depression.
In conclusion, this review is the first to evaluate the effects of early life inflammation on the development of future depression in a clinical setting. We show that increased inflammation in childhood may increase depression risk in later life, and although more robust clinical research is needed to definitively address the research questions raised by this article, we bring to light that timing may indeed matter. However, we should heed caution before classifying childhood as a potential ‘window of vulnerability’ owing to the vast limitations of the reviewed studies, and the insufficient research pertaining to the other two life stages. Being able to definitively pinpoint when in life individuals are at increased risk from an inflammatory challenge could (a) help practitioners and individuals better monitor and report these exposures/risk factors, (b) aid early intervention practices, through minimizing and efficiently treating inflammatory conditions during vulnerable stages of development, and (c) ultimately tailor treatment plans – all of which are much needed advancements in care practices.
Supplementary material
For supplementary material accompanying this paper visit http://dx.doi.org/10.1017/S0033291716000672.
Acknowledgements
This work was supported by the grant ‘Persistent Fatigue induced by interferon-alpha: a new immunological model for chronic fatigue syndrome’ (MR/J002739/1), and by the grant ‘Immuno-psychiatry: a consortium to test the opportunity for immunotherapeutics in psychiatry’ (MR/L014815/1), from the Medical Research Council (UK). Additional support was offered by the National Institute for Health Research Mental Health Biomedical Research Centre in Mental Health at South London and Maudsley NHS Foundation Trust and King’s College London. A.D.P. is supported by a studentship from Janssen Pharmaceutica, Johnson & Johnson.
Declaration of Interest
All authors receive research funding from companies interested in developing anti-inflammatory strategies for the treatment of depression. The data reviewed in this paper, however, are unrelated to this funding.
Open access
