The prospective association of adolescent loneliness and low resilience with anxiety and depression in young adulthood: The HUNT study

Abstract

Background

Loneliness is a recognized risk factor for anxiety and depression, yet research on its interaction with low resilience remains sparse, particularly across the adolescent-to-adult transition. This study investigates how adolescent loneliness, both independently and in interaction with low resilience, influences anxiety and depression in young adulthood.

Methods

This study utilized longitudinal data from The Trøndelag Health Study (HUNT) to track adolescents (13–19 years) from Young-HUNT3 (2006-08) through to HUNT4 (2017-19). Loneliness was assessed via a single-item measure, while resilience was estimated using the Resilience Scale for Adolescents. Anxiety and depression outcomes at the 11-year follow-up were measured using the Hospital Anxiety and Depression Scale. Multivariable logistic regression analyses were done to analyze the associations. Moreover, interaction effects were evaluated using relative excess risk due to interaction (RERI).

Results

Loneliness during adolescence independently predicted anxiety and depression in young adulthood. Adolescents experiencing both loneliness and low resilience showed notably higher risks compared to other groups (highly resilient adolescents without loneliness [reference], highly resilient adolescents with loneliness, and adolescents with low resilience and low loneliness). The combined effect of loneliness and low resilience exhibited a synergistic interaction on the additive scale, although it was not statistically significant (RERI 0.13, 95% CI -2.39-2.65).

Conclusion

Adolescent loneliness and low resilience independently predict anxiety and depression in young adulthood. The interaction between loneliness and low resilience further heightens these risks. This underscores the importance of early interventions that focus on resilience-building during adolescence and reducing the impacts of loneliness on mental health.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00127-025-02888-2.

Introduction

Mental health is fundamental to overall health, societal prosperity and global development [1, 2]. Anxiety and depression are the most debilitating mental health conditions worldwide, with over 300 million people affected by depression and around 45.82 million cases of anxiety disorders reported globally in 2019 [3–6]. These issues have been rising in Europe, particularly among adolescents and young adults, with around 13.6 million affected individuals in 2019, highlighting the need for global attention [7, 8]. The burden of mental illness in young adulthood often stems from factors experienced during adolescence, with loneliness being a significant, yet understudied factor [9, 10]. Through a deeper exploration of these dynamics, this study aims to illuminate the pathways linking adolescent loneliness to anxiety and depression in young adulthood.

Loneliness, characterized by a perceived lack of social connection compared to desired levels [11], regardless of social isolation status, can negatively affect individuals` life prospects. Research shows that loneliness represents a risk factor for depression [12, 13] and is also associated with anxiety [12, 14]. Research shows that loneliness can lead to a lack of social support and meaningful connections, further exacerbating feelings of distress and contributing to the onset of anxiety and depression [15]. While previous research has established the link between loneliness and mental health problems [16–18], there remains a gap in understanding the long-term effects of loneliness on mental health into adulthood [19]. Adolescent loneliness has gained public health attention recently due to its profound implications for individual and societal health [20, 21]. Adolescence is a phase of major biological, psychological, social, and cognitive transitions, making adolescents more vulnerable to loneliness than those in later life stages [22, 23], given their strive for independence from their families and the need to seek social and emotional connections with their peers [23]. A failure to form such intimate relationships during this developmental stage can heighten the risk of experiencing loneliness [24, 25]. The association between adolescent loneliness, anxiety and depression has been studied predominantly in cross-sectional studies [12, 26] and retrospectively [27, 28], but prospective studies with follow-up from childhood and adolescence to young adulthood are insufficient [29, 30]. Moreover, these limited prospective studies available have short follow-up periods, limiting the ability to capture long-term effects [18].

The impact of loneliness varies based on individual vulnerabilities, protective factors, and resources [31]. Resilience, a key protective factor, promotes healthy development by helping individuals cope with life’s challenges. Resilience refers to the process wherein an individual copes well to navigate life’s challenges by leveraging protective resources within themselves or their environment [32–35]. Resilience is a dynamic process, involving the ability to adapt and recover from adversity. It varies over time and situations, emphasizing the need for ongoing personal growth and support to cope with challenges [36]. Research shows resilience is negatively correlated with anxiety and depression,, suggesting it may protect against these conditions by buffering risk factors [37–43]. Low resilience, however, makes coping with stress harder, increasing vulnerability to mental health issues [40]. Despite extensive studies on childhood adversities, there is a gap in research on low resilience in adolescent loneliness and its impact on mental health. Understanding this relationship is key for effective interventions and promoting long-term well-being, especially in children and adolescents [44–46]. Therefore, we carried out a prospective cohort study of adolescents using The Trøndelag Health Study (HUNT) to (i) investigate the association of adolescent loneliness and low resilience with anxiety and depression in young adulthood and (ii) assess the joint association of adolescent loneliness and low resilience with anxiety and depression in young adulthood.

Materials and methods

Study population

The study used data from The Trøndelag Health Study (HUNT), a large population-based cohort in Trøndelag County, Norway [47–49]. The HUNT Study consists of four health surveys carried out at 11-year intervals, with Young-HUNT targeting adolescents (13–19 years) and HUNT for adults (20 + years) from the Trøndelag County in Norway. The Young-HUNT Study comprises three different waves: Young-HUNT1 (1995-97), Young-HUNT3 (2006-08), and Young-HUNT4 (2017-19) [47]. Adolescents from Young-HUNT3 (YH3) were included, with data collected via questionnaires, exams, and interviews. Adolescents not attending the school on the day of assessment received the questionnaire by mail [47]. The HUNT4 (2017-19) Survey took place 11 years later. The participants received an invitation letter and were asked to complete the electronic questionnaire and undertake physical tests at a local field station [49]. Of the 8199 adolescents who were included in YH3, 2293 (29%) attended HUNT4. Participation in the HUNT Study was voluntary, and all participants signed a written consent form. Written consent from parents or a guardian was required for participants under the age of 16 years. The Regional Committee for Medical and Health Research Ethics (Ref number: 286734) approved the present study.

Variables

Exposures variables at baseline; Young-HUNT3 (2006-08)

Loneliness

Loneliness was measured with the question, “Are you lonely?” on a five-point scale from " [1] Very rarely or never” to “ [5] Very often.“. The loneliness variable was grouped in two ways. Firstly, for the purpose of the multivariable logistic regression, categories 1 and 2 were grouped together to make a single category, “No,” with participants who reported low levels of loneliness. The other categories were kept unchanged. Secondly, the loneliness variable was dichotomized into the category Rarely lonely (‘Very rarely or never’ + ‘Rarely’ + ‘Sometimes lonely’) and the category Very lonely (‘Often’ + ‘Very often’) for the purposes of the interaction analyses. The dichotomization of the variable of loneliness has been previously utilized in studies focusing on adolescent populations [50–52]. The ‘Very lonely’ category was presumed to capture the most severe and high-risk cases of loneliness, and the ‘Rarely lonely’ category captured less severe and temporary forms or a complete absence of loneliness.

Resilience

Resilience factors were measured with 8 items from the Resilience Scale for Adolescents (READ) [53] in YH3, addressing social competence and family cohesion. Social competence items included items: I easily make others feel comfortable around me; I easily find new friends; I am good at talking to new people; I always find something fun to talk about and family cohesion: In my family we share views of what is important in life; I feel comfortable with my family; My family views the future as positive, even when sad things happen; In my family, we support each other. Responses ranged from [1]‘I totally disagree’ to [5]‘I totally agree’, with higher scores indicating higher resilience. The READ provides an exciting possibility to assess several different resilience factors with relatively few items, and can thus be used as a valuable measurement tool in resilience and risk factor research [54]. Furthermore, several studies support the use of the scale during all of adolescence, from 14 to 15 years of age up to 18 to 20 years of age. The scale may also be relevant in more applied settings, where it could be used to identify strengths and weaknesses in life areas that may enhance resilient outcomes in adolescents exposed to risk [54–56].

Demographics and baseline mental health

Covariables included age at baseline (13–15 and 15.1–19 years), gender, socioeconomic status (SES) at baseline, quality of relationship with family members, school satisfaction, and baseline self-esteem. Hopkins Symptom Checklist-25 (HSCL-25) was used as a control variable for the sensitivity analyses.

SES was assessed with the question: “How well off do you think your family is compared to most others?” Responses were: [1] ‘About the same as most others,’ [2] ‘Better,’ and [3] ‘Worse.’ The quality of family relationships was measured by: “How good is the relationship with your immediate family?” with responses from [1] ‘Very good’ to [4] ‘Bad.‘. School satisfaction was measured with nine questions on topics like concentration, arguing with teachers, school enjoyment, skipping school, understanding lessons, boredom, teasing, exclusion, and exam satisfaction, with responses from [1] ‘Never’ to [4] ‘Very often.‘. Self-esteem was addressed by the Rosenberg Self-esteem Scale [57]. This scale covered the questions: I have a positive attitude towards myself; I certainly feel useless at times; I feel I do not have much to be proud of; I feel that I’m a person of worth, at least on an equal plane with others. Responses ranged from [1] ‘I totally agree’ to [4] ‘I totally disagree’. The negatively phrased items on the school satisfaction and self-esteem scale were reverse-coded. Then, the scores were averaged and dichotomized. Scores from 1 to 2.99 were considered as low school satisfaction and self-esteem, while scores above 2.99 were considered as high. Age groups are displayed as 13–15 years and 16–19 years in tables. HSCL-5, a 5-item version of the HSCL-25, an extensively used self-report measure of anxiety and depression symptoms among adolescents, was used for sensitivity analyses. Compared with the HSCL-25, the short-form 5-item HSCL has acceptable validity [58]. The adolescents were asked if they had experienced each of the following during the last 14 days: Been constantly afraid and anxious, Felt tense or uneasy, Felt hopelessness about the future, Felt dejected or sad, Worried too much about various things. Each item was answered on a four-point scale: [1] ‘Not at all’ to [4] ‘Very much’ with the cut-off score ≥ 2.

Outcome variables; HUNT4 (2017-19)

Anxiety and depression

Symptoms of anxiety and depression were measured with the Hospital Anxiety and Depression Scale (HADS), which consists of 14 questions of which seven measure anxiety symptoms and seven measure depression symptoms, during the past week. Each question is answered on a scale of 0–3, giving one total score for anxiety (HADS-A, range 0–21) and one total score for depression (HADS-D, range 0–21). A total score from 0 to 7 indicates normal state, 8 to 10 indicates borderline anxiety or depression state and 11–21 indicates symptomatology of anxiety or depression. In the analysis stratified by resilience level, we combined the last two categories (score 8–21) to define the presence of symptoms (yes, no), i.e., a cut-off score ≥ 8. A total HADS score (HADS-T) consists of the sum of the HADS-A and the HADS-D scores, and the range is from 0 to 42 points. The cut-off indicating HADS-T symptomatology in this study was set at 19 points as recommended in the literature [59, 60]. Additional details of HADS are described elsewhere [61, 62]. The HADS was used to describe the presence of symptoms and is not a psychiatric diagnosis.

Statistical analyses

In the study sample of 2293 participants, we evaluated loneliness at baseline and the prospective outcome of anxiety and depression in relation to other baseline characteristics. We computed the mean baseline resilience scores by aggregating responses to the set of eight questionnaire items. We then divided participants into high and low-resilience groups, delineating low resilience at the threshold of ≤ 25th percentile (cut off: 3.625). The dichotomization of resilience items has been done before in previous studies among adolescents [63–65]. Categorical variables describing the study samples were reported as frequencies and percentages. To study the association of loneliness and resilience with anxiety and depression, we estimated odds ratios (OR) with 95% confidence intervals (CIs) using multivariable logistic regression. The analyses were conducted separately for each of the three outcomes, i.e., anxiety (HADS-A), depression (HADS-D), and combined measures of anxiety and depression (HADS-T) with the dichotomized exposures of loneliness and resilience. We reviewed the literature and created Directed Acyclic Graphs (DAGs) to select covariates that could be associated with loneliness (exposure) and anxiety and depression (outcome) in the regression analyses.

In Model I, we adjusted for age and gender. Model II was adjusted for age, gender, SES, quality of relationship with family members, school satisfaction, and self-esteem as potential confounding factors. In Model III, we additionally adjusted for resilience and loneliness as continuous variables. To assess the joint association of adolescent loneliness and low resilience with anxiety and depression in young adulthood, we generated four subgroups, i.e. participants with high resilience without loneliness (reference group), participants with high resilience and with loneliness (ORA), participants with low resilience without loneliness (ORB) and participants with low resilience with loneliness (ORAB). Logistic regression analyses were conducted to examine the relationship between adolescent resilience and anxiety and depression in young adulthood. The joint associations of loneliness and low resilience with anxiety and depression were estimated by using the relative excess risk due to interaction (RERI) with 95% CI. The RERI was calculated using an additive model: RERI = ORAB– ORA– ORB + 1 [66]. The additive model is used to test the interaction between two or more risk factors that together assert their influence on disease risk [67–69]. In brief, RERI > 0 and the lower limit of 95% CI > 0 suggest a synergistic effect of loneliness and low resilience on anxiety and depression [69]. Moreover, in addition to interaction on the additive scale, we assessed interaction on the multiplicative scale as it is always best to present both the additive and multiplicative measures of interaction [69].

Sensitivity analyses

We performed sensitivity analyses by classifying participants in two ways. First, those scoring ≥ 8 on HADS-A or HADS-D were considered to exhibit symptoms of both anxiety and depression (N = 1443) [70–73]. Second, participants who had utilized healthcare—reported psychiatrist visits, polyclinic treatments in the past year, or intake of antidepressants and anxiolytics in the past month—along with those scoring ≥ 8 on either HADS set, were also considered to exhibit symptoms (N = 806). To address reverse causality, we included only those adolescents who scored < 2 on the HSCL-5 in YH3 (free from anxiety and depression at baseline) in another sensitivity analysis. This group included 1826 adolescents (22.27%). We repeated all statistical analyses in this baseline symptom-free sample.

All analyses were done using R.

Results

Among the 2293 participants, the prevalence of loneliness was 8.2% (N = 189). During the 11-year follow-up, anxiety prevalence was 16.6% (N = 380) and depression prevalence was 6.1% (N = 141) (Table 1). Loneliness was more prevalent among girls, 16–19-year-olds, the mentally distressed (HSCL ≥ 2), those with low resilience, poor family relationships, low self-esteem, low school satisfaction, and low SES. Anxiety and depression at follow-up were higher in women, 23–25-year-olds, the mentally distressed, those with low resilience, poor family relationships, low self-esteem, low school satisfaction, and low SES.

Table 1.

Baseline (2006–2008) characteristics of the study sample stratified on loneliness and symptoms of anxiety and depression in the follow-up (2017–2019)

Variables measured at baseline	Total	Young-HUNT3 2006–2008 (Baseline)		HUNT4 2017–2019 (Follow-up)
		Loneliness		Anxiety symptoms		Depression symptoms
		No.	%	No.	%	No.	%
Study cohort	2293	189	8.2	380	16.6	141	6.1
Missing		159	6.9	844	36.8	844	36.8
Gender
Girls	1320	135	10.7	278	30	91	9.8
Boys	973	54	6.2	102	19.7	50	9.7
Age
13–15 years	857	55	7.1
16–19 years	1436	134	10
23–25 years	501			102	31.5	43	13.3
26–29 years	1792			278	24.7	98	8.7
Adolescent mental health
HSCL ≥ 2	434	135	33.5	132	46.8	52	18.5
HSCL < 2	1826	54	3.1	243	21.3	87	7.6
Resilience
Low	626	141	23.8	143	35	69	16.8
High	1607	47	3.1	226	22.7	68	6.8
Relationship with family
Bad	50	17	34.7	15	41.7	13	37.1
Good	2102	172	8.3	340	25.6	118	8.9
Self-esteem
Low	832	160	20.4	208	37.5	86	15.5
High	1419	29	2.2	166	19.2	53	6.1
School satisfaction
Not satisfied	448	100	22.7	100	37.6	49	18.4
Satisfied	1687	85	5.1	252	23.2	77	7.1
Socio-economic status
Low	179	46	26.1	44	37.6	20	16.8
Middle-class	334	22	6.7	66	30.1	20	9.1
High	1626	118	7.3	244	23.9	88	8.6

Adolescents within the loneliness variable category “Quite a lot” and “Very much” Hospital Anxiety and Depression scale ≥ 8

Associations of adolescent loneliness and anxiety and depression in young adulthood

We found an association between adolescent loneliness and anxiety and depression in young adulthood (Table 2). Participants with “Often” and “Very often” loneliness at baseline had an OR of 3.53 (95% CI 2.13–5.86) and 3.71 (95% CI 2.13–6.47) respectively for anxiety (HADS-A), after adjusting for age and gender. After adjustment for additional confounders (Model II), the OR was reduced to 2.52 (95% CI 1.47–4.34) and 2.42 (95% CI 1.32–4.45). Further adjustment for resilience (Model III) resulted in lower OR. Similar associations were noted for both depression (HADS-D) and the total measures of anxiety and depression (HADS-T).

Table 2.

The association of loneliness in adolescence with anxiety and depression in young adulthood

Loneliness	Total	N (%)	Model I OR (95% CI)	Model II OR (95% CI)	Model III OR (95% CI)
Anxiety (HADS-A)
No	906	182 (20)	Ref	Ref	Ref
Sometimes	325	111 (34.2)	2.0 (1.5–2.65)	1.54 (1.14–2.1)	1.51 (1.11–2.06)
Often	68	32 (47)	3.53 (2.13–5.86)	2.52 (1.47–4.34)	2.38 (1.37–4.13)
Very often	56	28 (50)	3.71 (2.13–6.47)	2.42 (1.32–4.45)	2.28 (1.23–4.23)
Depression (HADS-D)
No	905	63 (7)	Ref	Ref	Ref
Sometimes	325	41 (12.6)	1.99 (1.31–3.04)	1.34 (0.84–2.14)	1.26 (0.78–2.01)
Often	68	15 (22)	4.18 (2.21–7.92)	2.45 (1.2–4.99)	2.08 (1.01–4.29)
Very often	57	12 (21)	3.61 (1.8–7.25)	1.45 (0.64–3.3)	1.25 (0.54–2.87)
Anxiety and Depression (HADS T)
No	911	48 (5.3)	Ref	Ref	Ref
Sometimes	329	37 (11.2)	2.24 (1.42–3.53)	1.5 (0.9–2.49)	1.43 (0.86–2.38)
Often	68	11 (16.2)	3.71 (1.81–7.59)	2.39 (1.08–5.26)	2.08 (0.93–4.64)
Very often	57	12 (21.1)	4.44 (2.18–9.05)	2.25 (0.99–5.12)	1.98 (0.86–4.56)

Note CI, confidence interval; HADS-T, Total Hospital Anxiety and Depression scale; OR, odds ratio

HADS-A scale ≥ 8

HADS-D scale ≥ 8

HADS-T scale ≥ 19

Model I adjusted for age, gender

Model II adjusted for age, gender, relationship with family, SES, school satisfaction, and self-esteem at baseline. Model III additionally adjusted for resilience as a continuous variable

Associations of adolescent resilience and anxiety and depression in young adulthood

We found an association between adolescents’ low resilience and anxiety and depression in young adulthood (Table 3). Participants with low resilience at baseline had an OR of 1.79 (95% CI 1.39–2.31) for anxiety (HADS-A) after adjusting for age and gender. After adjustment for confounders (Model II), the OR was reduced to 1.23 (95% CI 0.91–1.66). Additional adjustment for loneliness (Model III) reduced the OR even more. Similar associations were noted for depression (HADS-D) and for total measures of anxiety and depression (HADS-T).

Table 3.

The association of resilience in adolescence with anxiety and depression in young adulthood

Resilience	Total	N (%)	Model I OR (95% CI)	Model II OR (95% CI)	Model III OR (95% CI)
Anxiety (HADS-A)
High	997	226 (22.7)	Ref	Ref	Ref
Low	410	143 (35)	1.79 (1.39–2.31)	1.23 (0.91–1.66)	1.07 (0.78–1.47)
Depression (HADS-D)
High	997	68 (6.8)	Ref	Ref	Ref
Low	410	69 (16.8)	2.85 (1.99–4.09)	1.73 (1.11–2.7)	1.57 (0.99–2.49)
Anxiety and Depression (HADS-T)
High	1004	54 (5.4)	Ref	Ref	Ref
Low	415	59 (14.2)	2.91 (1.96–4.3)	1.66 (1.03–2.7)	1.48 (0.9–2.44)

Model I adjusted for age, gender

Model II adjusted for age, gender, relationship with family, SES, school satisfaction, and self-esteem at baseline. Model III additionally adjusted for loneliness

Low resilience delineated at the threshold of ≤ 25th percentile (3.625)

Joint associations of adolescent loneliness and low resilience with anxiety and depression in young adulthood

Table 4 shows the joint association of adolescent loneliness and low resilience with anxiety and depression in young adulthood. For anxiety (HADS-A), lonely individuals with high resilience had an OR of 2.05 (95% CI: 0.96–4.35), low resilience without loneliness had an OR of 1.13 (95% CI: 0.81–1.58), and those with both had an OR of 2.11 (95% CI: 1.28–3.47). For depression (HADS-D), lonely individuals with high resilience had an OR of 2.36 (95% CI: 0.87–6.39), low resilience without loneliness had an OR of 1.81 (95% CI: 1.11–2.93), and those with both had an OR of 2.31 (95% CI: 1.17–4.55). For the total score (HADS-T), lonely individuals with high resilience had an OR of 1.87 (95% CI: 0.59–5.90), low resilience without loneliness had an OR of 1.57 (95% CI: 0.92–2.67), and those with both had an OR of 2.57 (95% CI: 1.26–5.23). We found a positive interaction on the additive scale for loneliness and resilience at baseline with the subsequent risk of anxiety and depression (HADS-T), with a RERI of 0.13 (95% CI -2.39-2.65), indicating the combined effect was greater than the sum of their individual effects, though not statistically significant. However, on the multiplicative scale, the interaction was negative for anxiety, depression, and the total score, meaning the combined effect was less than the product of their individual effects (Fig. 1).

Table 4.

Adjusted odds ratios for adolescent loneliness and low resilience at baseline associated with anxiety and depression in young adulthood: additive and multiplicative interaction

Loneliness	High resilience			Low Resilience			RERI (95% CI)¹
	Anxiety symptoms (HADS-A)
	Total	N	OR (95% CI)	Total	N	OR (95% CI)
Rarely lonely	925	201	1.00	296	91	1.13 (0.81–1.58)ORB	-0.07 (-1.83-1.69)
Very lonely	32	14	2.05 (0.96–4.35)ORA	92	46	2.11 (1.28–3.47)ORAB
Depression symptoms (HADS-D)
Rarely lonely	925	58	1.00	295	45	1.81 (1.11–2.93)ORB	-0.87 (-3.49-1.76)
Very lonely	32	7	2.36 (0.87–6.39)ORA	93	20	2.31 (1.17–4.55)ORAB
Anxiety and Depression (HADS-T)
Rarely lonely	931	48	1.00	299	36	1.57 (0.92–2.67)ORB	0.13 (-2.39-2.65)
Very lonely	32	4	1.87 (0.59–5.90)ORA	93	19	2.57 (1.26–5.23)ORAB

¹The relative excess risk due to interaction (RERI) between adolescent loneliness (A) and low resilience (B) was calculated using the formula: RERI = ORAB– ORA– ORB + 1

Adjusted for age, gender, relationship with family, SES, school satisfaction, and self-esteem at baseline

Measure of interaction on the multiplicative scale (HADS-A): 0.91 (95% CI. 0.37–2.23)

Measure of interaction on the multiplicative scale (HADS-D): 0.54 (95% CI 0.18–1.82)

Measure of interaction on the multiplicative scale (HADS-T): 0.88 (95% CI 0.26–3.62)

Fig. 1.

The risk of anxiety and depression in young adulthood associated with individual and joint exposures of adolescent loneliness and low resilience. Measurement of anxiety and depression levels in the study group, exclusively assessed using the Hospital Anxiety and Depression Scale Total (HADS-T). A HADS-T score of ≥ 19 indicates significant levels of anxiety and/or depression

Sensitivity analyses

The results of the sensitivity analyses and supplementary tables (S1-S7) are attached in separate files.

Discussion

In this study, adolescents reporting loneliness and low resilience showed an increased risk of anxiety and depression symptoms in young adulthood after an 11-year follow-up. The combination of loneliness and low resilience was synergistic and further elevated this risk in young adulthood. These findings align with previous research linking adolescent loneliness to higher anxiety and depression risk across all ages. For instance, Stickley et al. [26] reported higher odds of headaches, anxiety, and depressive symptoms among lonely adolescents. Moreover, a multilevel meta-analysis by Maes et al. [74] found a positive association between loneliness and social anxiety symptoms across childhood and adolescence. Additionally, Beutel et al. [12] highlighted a strong association between loneliness and generalized anxiety and depression, especially among middle and late adulthood individuals. Our findings align with studies by Kehusmaa et al. [29] and Xerxa et al. [30], linking childhood and adolescent loneliness with anxiety and depression in young adulthood. Moreover, sensitivity analyses are further supported by studies on healthcare utilization, such as Beutel et al. [12], who found lonely individuals more likely to seek treatment for anxiety and depression. Similarly, Junker et al. [75] showed that loneliness, often accompanied by anxiety and depression, increased self-harm hospitalization risk among Norwegian adolescents. Moreover, Modin et al. found that low childhood social status, reflecting loneliness, predicted higher hospitalization risk for anxiety and/or depression in girls [76]. The pathway between loneliness and anxiety and depression involves psychological, biological, and social factors [77, 78]. Loneliness can lead to negative cognitive patterns, such as rumination and self-criticism, which are linked to anxiety and depression [14, 79]. Moreover, the chronic stress associated with loneliness dysregulates biological systems, including the hypothalamic-pituitary-adrenal (HPA) axis and the immune system, which are crucial in the onset and persistence of anxiety and depression [80–82]. Loneliness affects youth’s diurnal cortisol, a key stress hormone [83, 84], potentially causing chronic increases in cortisol in young adulthood, leading to higher risks of depression, anxiety, inflammation, and other adverse health outcomes [85–87]. Adolescents’ heightened exposure to loneliness can be explained by their strive for independence from their families and seeking social and emotional connections with their peers [23]. Failure to form such intimate relationships during this developmental stage can increase the risk of experiencing loneliness [24, 25]. Chronic loneliness during adolescence can hinder social skill development, leading to low self-esteem and vulnerability and thus exacerbating anxiety and depression [88]. Moreover, loneliness disrupts cognitive processes, increasing susceptibility to anxiety and depression [89]. Lonelier individuals may exhibit disrupted emotional regulation, leading to increased negative emotions and a higher propensity for risky health behaviors, which could impact future mental health outcomes [90]. Additionally, lonelier individuals experience poorer sleep quality in young adulthood, potentially worsening their long-term mental health [91].

This study also examines low resilience as an additional risk factor in the pathway between adolescent loneliness, anxiety, and depression in young adulthood. Several cross-sectional studies across different age groups have highlighted resilience’s protective role. For instance, Zhao et al. found resilience mediated the link between loneliness and depression in the elderly, while studies on working-age adults and Norwegian adolescents showed resilience moderated this relationship [92–94]. In contrast to the protective effect, the detrimental effect of low resilience was shown by Fangyan et al. [95]. They compared depression levels among young adults with high and low resilience, showing that higher loneliness correlated with increased depression among those with low resilience, likely due to stress hormones like cortisol, which can lead to long-term neurological and physiological issues [96]. Moreover, individuals with low psychological resilience seem to have compromised control of brain circuits involved in emotion and fear [97] and poor cortico-limbic inhibition, representing a dysregulated emotional response to stress that may result in elevated cortisol exposure [98]. Likewise, loneliness and social isolation during adolescence precipitate a state of vulnerability or social frailty [80], rendering individuals more susceptible to adverse outcomes. This vulnerability may arise from a lack of protective factors like familial, social, or communal support, which is crucial for managing life’s pressures. Consequently, loneliness and isolation can diminish resilience, leaving individuals more vulnerable to anxiety and depression [99]. Our study builds on prior research with a large prospective cohort and an 11-year follow-up, adjusting for various covariates. While we found that adolescent loneliness and low resilience were both associated with an increased risk of anxiety and depression in adulthood, the evidence for a synergistic effect was weak. The interaction analyses yielded a small RERI value (0.13) with wide confidence intervals, suggesting uncertainty in the estimate. While this may indicate a potential additive interaction, the results should be interpreted carefully. The mechanisms underlying these associations remain complex, potentially involving compromised stress-coping mechanisms in lonely individuals with low resilience, leading to worsening social interactions and prolonged loneliness. Further research is needed to confirm causal relationships and to explore additional factors that may influence these dynamics.

Strengths and limitations

Our study’s strengths include a prospective design that establishes a clear temporal sequence, a large sample size at over an 11-year follow-up, and the ability to assess the interaction between adolescent loneliness and low resilience on young adult anxiety and depression. Additionally, the comprehensive dataset al.lowed control for various potential confounders. Resilience among adolescents was measured by the validated READ scale, covering relevant protective elements within resilience [54]. Similarly, anxiety and depression among young adults were measured by using the well-validated HADS questionnaire [61]. We conducted sensitivity analyses to ensure robustness, using an alternative definition of anxiety and depression that included those on antidepressant medication and those with recent healthcare visits, which increased precision and avoided misclassification. Additionally, we repeated all analyses after excluding adolescents who reported anxiety and depression at baseline.

This study has limitations. Self-reported measures may introduce biases such as social desirability and underreporting, while non-participation and absenteeism on data collection days may cause selection bias, potentially underestimating loneliness and its link to anxiety and depression. Despite an 11-year follow-up, more frequent assessments could better capture changes in loneliness and resilience. Self-reported loneliness is also prone to underreporting due to social stigma [100]. Loneliness is complex and ideally should be measured with a multi-item scale to account for variations in intensity, circumstances, and time with direct and indirect questions. Nevertheless, self-reported measures of loneliness have been shown to be equally reliable and valid as other forms of assessment [101]. Another issue is the categorization and applied dichotomization of loneliness and resilience. However, dichotomized outcomes may be necessary to communicate a comparison of risks [102]. Moreover, it is reassuring that this same dichotomization has been applied in earlier research among adolescents [50, 51, 103]. Moreover, despite the broad ranking and categorization of loneliness in this study, it fails to distinguish between adolescents who are rarely lonely and those who are never lonely. Another limitation is residual confounding despite controlling for relevant factors. Furthermore, it is uncertain if loneliness, low resilience, anxiety, and depression at baseline mutually influenced each other and thus increased the risk of anxiety and depression prospectively. Nevertheless, we excluded the adolescents who reported a score of ≥ 2 on the HSCL-5 at baseline in the sensitivity analyses to limit the possibility of reverse causation. In the current study, the HADS was used as a screening tool to capture symptoms of anxiety or depression and not a clinical diagnosis [104]. Although we strengthened the definition of anxiety and depression by adding medication use and hospital admission, we cannot rule out possible misclassification bias as we did not perform any objective tests. Thus, the limitations in identifying clinical anxiety or depression in adolescence and young adulthood, respectively, may limit the generalizability of our results.

We recognize the grade of attrition between the two surveys. However, this is a common issue in longitudinal studies [105, 106]. Several factors contributed to this: the long follow-up period (over 11 years), during which participants relocated or lost interest in follow-up studies, lower participation rates among young adults, as previously documented in HUNT research [47, 48], and relocation or death [107]. Moreover, the tendency of migration from rural Nord-Trøndelag to urban areas contributed to attrition [108, 109]. Mental health may have played a role, with those experiencing anxiety or depression being less likely to participate [110]. HUNT’s voluntary nature further challenges retention, and a lower response rate to the second questionnaire (Q2) in HUNT4 led to missing data on anxiety and depression (HADS) [49]. Despite this, strong baseline participation, consistent response patterns, and linkage to national registries support the study’s validity. Due to substantial missing data, we used complete case analysis rather than multiple imputation (MI) or weighting. MI assumes data are missing at random (MAR), but missingness in loneliness and anxiety/depression was likely due to participants experiencing these conditions, making MI unreliable [111]. Weighting does not fully remove bias when missingness relates to the outcome [111]. Non-responders were more likely to be older, male, or from vocational schools, suggesting socioeconomic differences [47], but these were already adjusted for in our models. Lastly, we adjusted for major confounders to account for potential bias due to missing data. But some level of bias might still be present.

Conclusion

In conclusion, our extensive cohort study indicates that adolescent loneliness and low resilience independently contribute to anxiety and depression in young adulthood. Low resilience interacting with loneliness can heighten this risk, reflecting the interconnectedness of the biological stress response. Our study underscores the importance of addressing loneliness and building resilience in adolescents to mitigate future anxiety and depression. Further research is needed to confirm the synergy between these risk factors and understand the mechanisms, potentially shifting resilience research toward a more holistic approach.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

All authors contributed substantially to the conceptualization and design of the study. NP was the main author, responsible for the overall structure and subject matter of the study and had primary responsibility for the layout of the work, attainment of data, data management, statistical analysis, and interpretation of data, drafting and revising the manuscript critically for important intellectual content. KK contributed to the design of the study, drafting, and revising the work critically for important intellectual content. LBS contributed to the design, revising the work critically for important intellectual content, and revision critically for important intellectual content. GAE contributed to revising the work critically for important intellectual content. UKM contributed to the acquisition of data, conceptualization, and design, and revising the work critically for important intellectual content. All authors approved the final version to be published and are accountable for all aspects of the work.

Funding

Open access funding provided by NTNU Norwegian University of Science and Technology (incl St. Olavs Hospital - Trondheim University Hospital)

Funding for open-access publication fees was received from the Norwegian University of Science and Technology (NTNU).

Data availability

The HUNT databank provided the data used in this study, but access is restricted. Data can be accessed by approaching the author and permission from HUNT, The Regional Ethical Committee, and the Norwegian Data Protection Authority. All the subjects of this study are stored in the HUNT data using a personal identification number as an I.D. The HUNT Research Centre stores and uses these data with authorization from the Norwegian Data Inspectorate without breaking participant privacy. The researcher will always receive an anonymous dataset after approval from the Regional Ethical Committee and HUNT Research Centre. For more information about HUNT data see https://www.ntnu.edu/hunt/data.

Declarations

Ethical approval

The HUNT Study was approved by the Data Inspectorate of Norway and recommended by the Regional Committee for Ethics in Medical Research. The ethical approval for conducting this study was also obtained from the Regional Committee for Ethics in Medical Research (REK Nord; reference number 286734).

Consent for publication

Not applicable.

Consent to participate

Informed consent was obtained from all individual participants of all the cohorts included in this study, as well as from legal guardians/next of kin if the student was under 16 years of age.

Competing interests

The authors declare no competing interests.