Abstract
Background
Adverse perinatal exposures have been associated with psychiatric disorders and suicidal behaviours later in life. However, the independent associations of gestational age at birth or fetal growth with suicide death, potential sex-specific differences, and causality of these associations are unclear.
Methods
A national cohort study was conducted of all 2 440 518 singletons born in Sweden during 1973–98 who survived to age 18 years, who were followed up through 2016. Cox regression was used to compute hazard ratios (HRs) for suicide death associated with gestational age at birth or fetal growth while mutually adjusting for these factors, sociodemographic characteristics and family history of suicide. Co-sibling analyses assessed the influence of unmeasured shared familial (genetic and/or environmental) factors.
Results
In 31.2 million person-years of follow-up, 4470 (0.2%) deaths by suicide were identified. Early preterm birth (22–33 weeks) was associated with an increased risk of suicide among females [adjusted hazard ratio (HR), 1.97; 95% confidence interval CI), 1.29, 3.01; P = 0.002) but not males (0.90; 0.64, 1.28; P = 0.56), compared with full-term birth (39–41 weeks). Small for gestational age was associated with a modestly increased risk of suicide among females (adjusted HR, 1.27; 95% CI, 1.08, 1.51; P = 0.005) and males (1.14; 1.03, 1.27; P = 0.02). However, these associations were attenuated and non-significant after controlling for shared familial factors.
Conclusions
In this large national cohort, preterm birth in females and low fetal growth in males and females were associated with increased risks of suicide death in adulthood. However, these associations appeared to be non-causal and related to shared genetic or prenatal environmental factors within families.
Keywords: Fetal growth, gestational age, premature birth, small for gestational age, suicide
Key Messages
Adverse perinatal exposures have been associated with suicidal behaviours later in life; however, the independent associations of gestational age at birth or fetal growth with suicide death, potential sex-specific differences and causality are unclear.
In a large national cohort study, preterm birth in females and low fetal growth in males and females were associated with increased risks of suicide death in adulthood.
In co-sibling analyses, these associations appeared to be non-causal and related to shared genetic or prenatal environmental factors within families.
Further elucidation of common exposures that lead to preterm birth or low fetal growth and future suicidality is needed to develop a life course model for suicide risk assessment.
Introduction
Suicide is a major public health problem and accounts for 800 000 deaths worldwide each year.1,2 Most research into suicide risk pathways has focused on psychiatric disorders and other proximal risk factors in adolescence or adulthood.3,4 However, adverse perinatal exposures, including preterm birth and low fetal growth, also have been associated with psychiatric and other chronic disorders later in life,5–7 consistent with the developmental origins of health and disease hypothesis.8 Such evidence has prompted investigations into whether these exposures also are associated with higher risk of suicide. A better understanding of perinatal risk factors is needed to develop a life course model for suicide risk assessment and potentially may enable earlier interventions in high-risk subgroups.
A recent meta-analysis reported that low birthweight and small for gestational age (SGA) were associated with 1.2- to 1.3-fold risks of suicide later in life.9 However, the independent effects of gestational age at birth and fetal growth, potential sex-specific differences and the causal nature of these associations remain unclear. Preterm birth and low fetal growth have different underlying aetiologies and may have different long-term sequelae that require further delineation. Death by suicide is a leading cause of premature mortality among men and women, but is 2 to 4 times more common in men.1,2 Most previous studies have had insufficient numbers or statistical power to examine perinatal risk factors for suicide specifically in women. In addition, it is unclear whether previously reported associations are causal or related to shared genetic or environmental factors in the families of those affected. Large population-based cohorts are needed to enable well-powered analyses in subgroups and family-based designs to explore potential causality.
We sought to address these knowledge gaps by conducting a national cohort study of 2.4 million adults in Sweden. Our goals were to: (i) provide population-based risk estimates for suicide associated with gestational age at birth and fetal growth in the largest cohort to date that was followed up into adulthood; (ii) assess for sex-specific differences; and (iii) explore the influence of shared familial (genetic and/or environmental) factors using co-sibling analyses. We hypothesized that perinatal factors are associated with increased risk of suicide in adult men and women, and that these associations are partially explained by shared familial confounders.
Methods
This study was approved by the Regional Ethical Review Board in Lund, Sweden (No. 2016/679) and was conducted in accordance with the Declaration of Helsinki.
Study population
The Swedish Medical Birth Register contains prenatal and birth information for nearly all births in Sweden since 1973.10 Using this register, we identified 2 447 526 persons who were born as singletons during 1973–98 and who survived to age 18 years. Singleton births were chosen to improve internal comparability, given the higher prevalence and different underlying causes of preterm birth and low birthweight among multiple births.11 We excluded 7008 (0.3%) births that had missing information for gestational age, leaving 2 440 518 births (99.7% of the original cohort) for inclusion in the study.
Ascertainment of gestational age at birth and fetal growth
Gestational age at birth was identified from the Swedish Medical Birth Register based on maternal report of last menstrual period in the 1970s and ultrasound estimation starting in the 1980s and later. This was examined alternatively as a continuous variable or categorical variable with five groups: early preterm (22–33 weeks), late preterm (34–36 weeks), early term (37–38 weeks), full-term (39–41 weeks, used as the reference group) and post-term (≥42 weeks). In addition, the first two groups were combined to provide summary estimates for preterm birth (<37 weeks). Only 0.2% of all births were extremely preterm (22–27 weeks), which were too few to allow separate analysis of this subgroup because of the low number of suicides.
Fetal growth was based on standardized measures of birthweight for gestational age in the Swedish Medical Birth Register and classified in two ways: (i) small for gestational age (SGA; defined as birthweight <10th percentile for gestational age), appropriate for gestational age (AGA; 10th to 90th percentile), or large for gestational age (LGA; >90th percentile); and (ii) number of standard deviations (SD) from the mean birthweight for gestational age and sex, in five groups (<-2, -2 to -1, -1 to 1, 1 to 2, >2 SD).
Suicide ascertainment
The study cohort was followed up for suicide death from age 18 years through 31 December 2016 (maximum age 44.0 years; median 30.0) using nationwide data from the Swedish Cause of Death Register. This register includes deaths and International Classification of Diseases (ICD) codes for cause of death among all persons registered in Sweden since 1960, with compulsory reporting nationwide. All intentional deaths were identified using ICD-10 codes X60-X84, and deaths of undetermined intent were identified using ICD-10 codes Y10-Y34. Previous studies have indicated that substantial numbers of suicides are misclassified as deaths of undetermined intent.12,13 In the present study, intentional deaths and deaths of undetermined intent were analysed together as the primary outcome and separately as secondary outcomes in a sensitivity analysis.
Covariates
Other perinatal and sociodemographic characteristics that may be associated with gestational age at birth or fetal growth and suicide were identified using the Swedish Medical Birth Register, Cause of Death Register and national census data, which were linked using a pseudonymous personal identification number. Covariates included birth year (modelled simultaneously as a continuous variable and categorical variable by decade), sex, birth order (1, 2, ≥3), maternal and paternal age at birth of the index offspring (5-year intervals), highest maternal and paternal education level achieved (≤9, 10–12, >12 years) and family history of suicide in a first-degree relative (yes/no). All variables were >98.5% complete. Missing data were modelled as a separate category using a missing data indicator, thus allowing all individuals to be retained in the analyses, and had a negligible effect on the results due to their rarity.
Statistical analysis
Cox proportional hazards regression was used to compute hazard ratios (HRs) and 95% confidence intervals (CIs) for associations between gestational age at birth or fetal growth and risk of suicide death. Analyses were conducted both unadjusted and adjusted for covariates (as above). Sex-specific differences were assessed by performing sex-stratified analyses and formally testing for interactions between gestational age at birth or fetal growth and sex on the additive and multiplicative scale.14 The proportional hazards assumption was assessed by examining log-log plots, which showed no substantial departures.
Co-sibling analyses were performed to assess the potential influence of unmeasured shared familial (genetic and/or environmental) factors on the observed associations.7 Relevant exposures that may be shared among siblings include maternal psychosocial stress, prenatal substance use and parental psychopathology. A total of 2 072 861 (84.9%) individuals had at least one sibling and were included in these analyses. Stratified Cox regression models were performed using a separate stratum for each family as identified by the mother’s anonymous identification number. In these models, each set of siblings had its own baseline hazard function that reflected their shared genetic and environmental factors, thus controlling for their shared exposures even if not specifically measured. In addition, these analyses were further adjusted for the same covariates as in the main analyses.
In a secondary analysis, we explored interactions between preterm birth and fetal growth in relation to suicide on the additive and multiplicative scale.14 This analysis examined whether the association between low fetal growth and suicide risk varies according to gestational age at birth, or vice versa, which to our knowledge has not been formally assessed. All statistical tests were two-sided and used an α-level of 0.05. All analyses were conducted using Stata version 15.1.
Results
Table 1 reports offspring and parental characteristics in the total cohort, in offspring born preterm or SGA and in those who died by suicide. Persons born preterm or SGA were more likely to be male or firstborn, and their parents were more likely to be <25 years old or have low education level. Those who died by suicide were more likely to be male or have a family history of suicide, and their parents also were more likely to be <25 years old or have low education level.
Table 1.
Characteristics of study participants (1973–2016), Sweden
| Total cohort | Born preterm | Born SGA | Died by suicide | |
|---|---|---|---|---|
| N = 2 440 518 (100.0%) | N = 120 700 (4.9%) | N = 244 088 (10.0%) | N = 4470 (0.2%) | |
| n (%) | n (%) | n (%) | n (%) | |
| Offspring factors | ||||
| Sex | ||||
| Male | 1253571 (51.4) | 66102 (54.8) | 126405 (51.8) | 3270 (73.2) |
| Female | 1186947 (48.6) | 54598 (45.2) | 117683 (48.2) | 1200 (26.8) |
| Gestational age at birth (weeks) | ||||
| Preterm (<37) | 120700 (4.9) | 120700 (100.0) | 12773 (5.2) | 236 (5.3) |
| Early preterm (22-33) | 28024 (1.1) | 28024 (23.2) | 3550 (1.4) | 54 (1.2) |
| Late preterm (34-36) | 92676 (3.8) | 92676 (76.8) | 9223 (3.8) | 182 (4.1) |
| Early term (37-38) | 414949 (17.0) | 0 (0.0) | 29299 (12.0) | 755 (16.9) |
| Full-term (39-41) | 1685349 (69.1) | 0 (0.0) | 157182 (64.4) | 3025 (67.7) |
| Post-term (≥42) | 219520 (9.0) | 0 (0.0) | 44834 (18.4) | 454 (10.2) |
| Fetal growth | ||||
| SGA (<10th percentile) | 244088 (10.0) | 12773 (10.6) | 244088 (100.0) | 573 (12.8) |
| AGA (10th to 90th percentile) | 1952395 (80.0) | 98171 (81.3) | 0 (0.0) | 3487 (78.0) |
| LGA (>90th percentile) | 244035 (10.0) | 9756 (8.1) | 0 (0.0) | 410 (9.2) |
| <-2 SD from mean | 51664 (2.1) | 3436 (2.8) | 51664 (21.2) | 140 (3.1) |
| -2 to -1 SD from mean | 309903 (12.7) | 14556 (12.1) | 192424 (78.8) | 666 (14.9) |
| -1 to 1 SD from mean | 1715140 (70.3) | 88603 (73.4) | 0 (0.0) | 3066 (68.6) |
| 1 to 2 SD from mean | 298661 (12.2) | 10572 (8.8) | 0 (0.0) | 487 (10.9) |
| >2 SD from mean | 65150 (2.7) | 3533 (2.9) | 0 (0.0) | 111 (2.5) |
| Birth order | ||||
| 1 | 1018391 (41.7) | 59177 (49.0) | 134574 (55.1) | 1801 (40.3) |
| 2 | 897443 (36.8) | 35585 (29.5) | 71074 (29.1) | 1610 (36.0) |
| ≥3 | 524684 (21.5) | 25938 (21.5) | 38440 (15.8) | 1059 (23.7) |
| Family history of suicide | 21489 (0.9) | 1206 (1.0) | 2479 (1.0) | 157 (3.5) |
| Maternal factors | ||||
| Age (years) | ||||
| <20 | 91438 (3.7) | 6248 (5.2) | 13467 (5.5) | 330 (7.4) |
| 20-24 | 600682 (24.6) | 30361 (25.1) | 70359 (28.8) | 1373 (30.7) |
| 25-29 | 906639 (37.1) | 40143 (33.3) | 86896 (35.6) | 1451 (32.5) |
| 30-34 | 587489 (24.1) | 27865 (23.1) | 50752 (20.8) | 898 (20.1) |
| 35-39 | 211183 (8.7) | 12680 (10.5) | 18408 (7.5) | 335 (7.5) |
| ≥40 | 38380 (1.6) | 2964 (2.5) | 3568 (1.5) | 80 (1.8) |
| Unknown | 4707 (0.2) | 439 (0.4) | 638 (0.3) | 3 (0.1) |
| Mean ± SD | 28.2 ± 5.1 | 28.3 ± 5.6 | 27.5 ± 5.2 | 27.2 ± 5.5 |
| Median | 27.8 | 27.9 | 27.1 | 26.7 |
| Education (years) | ||||
| ≤9 | 353470 (14.5) | 20920 (17.3) | 45069 (18.5) | 944 (21.1) |
| 10-12 | 1219721 (50.0) | 62214 (51.5) | 125751 (51.5) | 2264 (50.7) |
| >12 | 864293 (35.4) | 37216 (30.8) | 72807 (29.8) | 1258 (28.1) |
| Unknown | 3034 (0.1) | 350 (0.3) | 461 (0.2) | 4 (0.1) |
| Paternal factors | ||||
| Age (years) | ||||
| <25 | 335669 (13.7) | 18380 (15.2) | 42188 (17.3) | 917 (20.5) |
| 25-29 | 806719 (33.1) | 36152 (30.0) | 82297 (33.7) | 1428 (31.9) |
| 29-34 | 729010 (29.9) | 31630 (26.2) | 65604 (26.9) | 1135 (25.4) |
| 35-39 | 356982 (14.6) | 17317 (14.3) | 30601 (12.5) | 597 (13.4) |
| 40-44 | 128567 (5.3) | 7158 (5.9) | 11607 (4.8) | 240 (5.4) |
| ≥45 | 54502 (2.2) | 3361 (2.8) | 5484 (2.2) | 115 (2.6) |
| Unknown | 29069 (1.2) | 6702 (5.6) | 6307 (2.6) | 38 (0.9) |
| Mean ± SD | 31.0 ± 5.9 | 31.1 ± 6.4 | 30.5 ± 6.1 | 30.4 ± 6.4 |
| Median | 30.3 | 30.3 | 29.7 | 29.6 |
| Education (years) | ||||
| ≤9 | 557325 (22.8) | 28693 (23.8) | 63866 (26.2) | 1361 (30.4) |
| 10-12 | 1178917 (48.3) | 56747 (47.0) | 116194 (47.6) | 2022 (45.2) |
| >12 | 676117 (27.7) | 28592 (23.7) | 57785 (23.7) | 1042 (23.3) |
| Unknown | 28159 (1.2) | 6668 (5.5) | 6243 (2.6) | 45 (1.0) |
AGA, appropriate for gestational age; LGA, large for gestational age; SD, standard deviation; SGA, small for gestational age.
A total of 4470 persons (0.2% of the cohort) died by suicide. The median age for the entire cohort at the end of follow-up was 30.0 years (mean 30.9 ± 7.4) and the median age at death by suicide was 25.9 years (mean 27.0 ± 5.7).
Gestational age at birth or fetal growth and risk of suicide
Table 2 shows associations between gestational age at birth or fetal growth and risk of suicide. The prevalence of suicide death was 0.20% among all persons born preterm vs 0.18% among those born at full term, and 0.23% among those born SGA vs 0.18% among those born AGA. In the total cohort, preterm birth was associated with a modestly increased risk of suicide before but not after adjusting for covariates (unadjusted HR, 1.15; 95% CI, 1.01, 1.32; adjusted HR, 1.07; 0.94, 1.22). Low fetal growth was associated with a modestly increased risk of suicide even after adjusting for covariates (e.g. adjusted HR, SGA vs AGA: 1.18; 95% CI, 1.08, 1.29; P < 0.001; <-2 vs -1 to 1 SD: 1.27; 1.07, 1.50; P = 0.007). Across the full range of fetal growth, each additional 1 SD was associated with a 7% lower risk of suicide on average (adjusted HR, 0.93; 0.90, 0.96; P < 0.001). Mutual adjustment for gestational age at birth and fetal growth had a negligible effect on suicide risk estimates associated with either of these factors.
Table 2.
Associations between gestational age at birth or fetal growth (1973–98) and risk of suicide death (1991–2016), Sweden
| No. suicides | Prev.a | Unadjusted |
Adjustedb |
||
|---|---|---|---|---|---|
| HR (95% CI) | HR (95% CI) | P | |||
| Total cohort | |||||
| Gestational age at birth (weeks) | |||||
| Preterm (<37) | 236 | 0.20% | 1.15 (1.01, 1.32) | 1.07 (0.94, 1.22) | 0.31 |
| Early preterm (22-33) | 54 | 0.19% | 1.27 (0.97, 1.66) | 1.16 (0.89, 1.52) | 0.28 |
| Late preterm (34-36) | 182 | 0.20% | 1.12 (0.97, 1.31) | 1.05 (0.90, 1.22) | 0.54 |
| Early term (37-38) | 755 | 0.18% | 1.05 (0.97, 1.14) | 1.02 (0.95, 1.11) | 0.56 |
| Full-term (39-41) | 3,025 | 0.18% | Reference | Reference | |
| Post-term (≥42) | 454 | 0.21% | 1.03 (0.93, 1.13) | 0.99 (0.89, 1.09) | 0.82 |
| Per additional 1 week (trend) | 0.98 (0.97, 1.00) | 0.99 (0.97, 1.01) | 0.25 | ||
| Fetal growth (percentile) | |||||
| SGA (<10th) | 573 | 0.23% | 1.19 (1.09, 1.30) | 1.18 (1.08, 1.29) | <0.001 |
| AGA (10th to 90th) | 3,487 | 0.18% | Reference | Reference | |
| LGA (>90th) | 410 | 0.17% | 0.99 (0.90, 1.10) | 0.95 (0.86, 1.05) | 0.33 |
| Fetal growth (SD from mean) | |||||
| <-2 | 140 | 0.27% | 1.32 (1.12, 1.57) | 1.27 (1.07, 1.50) | 0.007 |
| -2 to -1 | 666 | 0.21% | 1.11 (1.02, 1.21) | 1.12 (1.02, 1.21) | 0.01 |
| -1 to 1 | 3.066 | 0.18% | Reference | Reference | |
| 1 to 2 | 487 | 0.16% | 0.96 (0.87, 1.05) | 0.93 (0.84, 1.02) | 0.12 |
| >2 | 111 | 0.17% | 1.02 (0.84, 1.23) | 0.94 (0.78, 1.14) | 0.56 |
| Per additional 1 SD (trend) | 0.94 (0.91, 0.97) | 0.93 (0.90, 0.96) | <0.001 | ||
| Males | |||||
| Gestational age at birth (weeks) | |||||
| Preterm (<37) | 165 | 0.25% | 1.02 (0.87, 1.20) | 0.99 (0.84, 1.15) | 0.85 |
| Early preterm (22-33) | 32 | 0.21% | 0.95 (0.67, 1.35) | 0.90 (0.64, 1.28) | 0.56 |
| Late preterm (34-36) | 133 | 0.26% | 1.04 (0.87, 1.24) | 1.01 (0.85, 1.20) | 0.93 |
| Early term (37-38) | 567 | 0.26% | 1.05 (0.96, 1.15) | 1.03 (0.94, 1.13) | 0.48 |
| Full-term (39-41) | 2,195 | 0.26% | Reference | Reference | |
| Post-term (≥42) | 343 | 0.30% | 1.06 (0.94, 1.18) | 1.03 (0.92, 1.15) | 0.64 |
| Per additional 1 week (trend) | 1.00 (0.99, 1.02) | 1.01 (0.99, 1.03) | 0.49 | ||
| Fetal growth (percentile) | |||||
| SGA (<10th) | 410 | 0.32% | 1.15 (1.04, 1.28) | 1.14 (1.03, 1.27) | 0.02 |
| AGA (10th to 90th) | 2,559 | 0.26% | Reference | Reference | |
| LGA (>90th) | 301 | 0.23% | 0.92 (0.82, 1.04) | 0.91 (0.81, 1.02) | 0.12 |
| Fetal growth (SD from mean) | |||||
| <-2 | 106 | 0.37% | 1.27 (1.05, 1.55) | 1.25 (1.03, 1.53) | 0.03 |
| −2 to -1 | 462 | 0.30% | 1.06 (0.96, 1.18) | 1.06 (0.96, 1.17) | 0.27 |
| −1 to 1 | 2,264 | 0.26% | Reference | Reference | |
| 1 to 2 | 354 | 0.22% | 0.89 (0.79, 0.99) | 0.88 (0.78, 0.98) | 0.02 |
| >2 | 84 | 0.23% | 0.94 (0.76, 1.17) | 0.92 (0.74, 1.14) | 0.43 |
| Per additional 1 SD (trend) | 0.93 (0.90, 0.96) | 0.93 (0.90, 0.96) | <0.001 | ||
| Females | |||||
| Gestational age at birth (weeks) | |||||
| Preterm (<37) | 71 | 0.13% | 1.39 (1.09, 1.77) | 1.34 (1.05, 1.71) | 0.02 |
| Early preterm (22-33) | 22 | 0.17% | 2.06 (1.35, 3.15) | 1.97 (1.29, 3.01) | 0.002 |
| Late preterm (34-36) | 49 | 0.12% | 1.21 (0.91, 1.62) | 1.17 (0.88, 1.56) | 0.29 |
| Early term (37-38) | 188 | 0.09% | 1.00 (0.85, 1.17) | 0.99 (0.85, 1.17) | 0.94 |
| Full-term (39-41) | 830 | 0.10% | Reference | Reference | |
| Post-term (≥42) | 111 | 0.11% | 0.93 (0.76, 1.13) | 0.89 (0.73, 1.08) | 0.24 |
| Per additional 1 week (trend) | 0.95 (0.92, 0.98) | 0.95 (0.92, 0.98) | 0.001 | ||
| Fetal growth (percentile) | |||||
| SGA (<10th) | 163 | 0.14% | 1.29 (1.09, 1.52) | 1.27 (1.08, 1.51) | 0.005 |
| AGA (10th to 90th) | 928 | 0.10% | Reference | Reference | |
| LGA (>90th) | 109 | 0.10% | 1.08 (0.89, 1.32) | 1.08 (0.88, 1.32) | 0.46 |
| Fetal growth (SD from mean) | |||||
| <-2 | 34 | 0.15% | 1.32 (0.94, 1.87) | 1.30 (0.92, 1.84) | 0.13 |
| -2 to -1 | 204 | 0.13% | 1.28 (1.09, 1.49) | 1.27 (1.09, 1.48) | 0.003 |
| -1 to 1 | 802 | 0.10% | Reference | Reference | |
| 1 to 2 | 133 | 0.10% | 1.07 (0.89, 1.28) | 1.07 (0.89, 1.29) | 0.44 |
| >2 | 27 | 0.09% | 1.06 (0.72, 1.56) | 1.03 (0.70, 1.52) | 0.86 |
| Per additional 1 SD (trend) | 0.93 (0.88, 0.99) | 0.94 (0.88, 0.99) | 0.03 | ||
AGA, appropriate for gestational age; CI, confidence interval; HR, hazard ratio; LGA, large for gestational age; SD, standard deviation; SGA, small for gestational age.
Cumulative prevalence of suicide.
The adjusted model included birth year, sex, gestational age at birth, fetal growth, birth order, maternal and paternal age and education and family history of suicide.
In sex-stratified analyses, preterm birth was associated with an increased risk of suicide among females (adjusted HR, 1.34; 95% CI, 1.05, 1.71; P = 0.02) but not males (0.99; 0.84, 1.15; P = 0.85), compared with full-term birth (Table 2). Among females, this increased risk was entirely due to an association with early preterm birth (22–33 weeks) (adjusted HR, 1.97; 95% CI, 1.29, 3.01; P = 0.002), whereas no association was found with late preterm birth (34–36 weeks). In contrast, low fetal growth was associated with a modestly increased risk of suicide among both males (adjusted HR, SGA vs AGA: 1.14; 95% CI, 1.03, 1.27; P = 0.02) and females (1.27; 1.08, 1.51; P = 0.005). Across the full range of fetal growth, each additional 1 SD was associated with a 7% lower risk of suicide among males (adjusted HR, 0.93; 95% CI, 0.90, 0.96; P < 0.001) and a 6% lower risk among females (0.94; 0.88, 0.99; P = 0.03).
Interactions between preterm birth or fetal growth and sex are shown in Supplementary Tables S1 and S2, available as Supplementary data at IJE online. Male sex and preterm birth had a significant negative interaction on the multiplicative but not additive scale (i.e. their combined effect on suicide risk was less than the product of their separate effects, P = 0.02; Supplementary Table S1,). No interactions were found between sex and fetal growth (Supplementary Table S2). The absence of additive interactions suggests that preterm birth and SGA accounted for a similar number of suicides among men and women. Figures 1 and 2 show sex-stratified HRs for suicide associated with gestational age at birth or fetal growth, respectively.
Figure 1.
Sex-stratified hazard ratios for suicide death (1991–2016) by gestational age at birth compared with full-term (1973–98), Sweden.
Figure 2.
Sex-stratified hazard ratios for suicide death (1991–2016) by fetal growth relative to the mean (1973–98), Sweden.
Co-sibling analyses
After controlling for unmeasured shared familial factors in co-sibling analyses, all associations observed in the primary analyses were attenuated and non-significant (Table 3). For example, comparing SGA vs AGA births, the adjusted HRs for suicide in the entire cohort were 1.18 (95% CI, 1.08, 1.29; P < 0.001) in the primary analysis vs 1.08 (0.88, 1.32; P = 0.45) in the co-sibling analysis, among males were 1.14 (1.03, 1.27; P = 0.02) vs 0.88 (0.65, 1.20; P = 0.43), and among females were 1.27 (1.08, 1.51; P = 0.005) vs 0.97 (0.58, 1.60; P = 0.89). Among females, the adjusted HR for suicide associated with early preterm birth was 1.97 (95% CI, 1.29, 3.01; P = 0.002) in the primary analysis vs 0.98 (0.10, 9.41; P = 0.99) in the co-sibling analysis. The co-sibling results should be interpreted with caution because they had lower statistical power and precision than the primary analyses. However, they suggest that the associations observed between preterm birth or low fetal growth and suicide in the entire cohort were largely attributable to confounding by shared familial (genetic and/or prenatal environmental) factors.
Table 3.
Co-sibling analyses for associations between gestational age at birth or fetal growth (1973–98) and risk of suicide death (1991–2016), Sweden
| All |
Males |
Females |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| No. suicides | HR (95% CI)a | P | No. suicides | HR (95% CI)a | P | No. suicides | HR (95% CI)a | P | |
| Gestational age at birth (wks) | |||||||||
| Preterm (<37) | 190 | 1.06 (0.79 1.42) | 0.69 | 135 | 1.19 (0.77 1.84) | 0.42 | 55 | 0.53 (0.19 1.49) | 0.23 |
| Early preterm (22-33) | 42 | 1.31 (0.68 2.53) | 0.42 | 25 | 1.63 (0.60 4.40) | 0.34 | 17 | 0.98 (0.10 9.41) | 0.99 |
| Late preterm (34-36) | 148 | 1.02 (0.75 1.39) | 0.91 | 110 | 1.14 (0.72 1.79) | 0.58 | 38 | 0.48 (0.16 1.44) | 0.19 |
| Early term (37-38) | 586 | 1.08 (0.92 1.28) | 0.34 | 444 | 1.27 (0.99 1.62) | 0.06 | 142 | 0.82 (0.49 1.37) | 0.45 |
| Full-term (39-41) | 2426 | Reference | 1781 | Reference | 645 | Reference | |||
| Post-term (≥42) | 352 | 1.17 (0.95 1.44) | 0.14 | 264 | 1.17 (0.86 1.60) | 0.31 | 88 | 1.22 (0.71 2.10) | 0.46 |
| Per additional 1 week (trend) | 1.01 (0.97 1.05) | 0.61 | 1.00 (0.95 1.07) | 0.91 | 1.05 (0.93 1.19) | 0.41 | |||
| Fetal growth (percentile) | |||||||||
| SGA (<10th) | 415 | 1.08 (0.88 1.32) | 0.45 | 296 | 0.88 (0.65 1.20) | 0.43 | 119 | 0.97 (0.58 1.60) | 0.89 |
| AGA (10th to 90th) | 2804 | Reference | 2083 | Reference | 721 | Reference | |||
| LGA (>90th) | 335 | 1.10 (0.88 1.37) | 0.41 | 245 | 1.14 (0.82 1.57) | 0.44 | 90 | 1.23 (0.72 2.11) | 0.45 |
| Fetal growth (SD from mean) | |||||||||
| <-2 | 104 | 0.95 (0.64 1.40) | 0.80 | 79 | 0.74 (0.39 1.39) | 0.35 | 25 | 2.03 (0.64 6.44) | 0.23 |
| -2 to -1 | 487 | 1.17 (0.97 1.40) | 0.10 | 343 | 1.03 (0.78 1.37) | 0.84 | 144 | 1.06 (0.64 1.74) | 0.82 |
| -1 to 1 | 2474 | Reference | 1845 | Reference | 629 | Reference | |||
| 1 to 2 | 398 | 1.10 (0.90 1.34) | 0.34 | 290 | 1.15 (0.86 1.54) | 0.35 | 108 | 1.12 (0.69 1.85) | 0.64 |
| >2 | 91 | 1.13 (0.77 1.66) | 0.54 | 67 | 0.79 (0.42 1.50) | 0.48 | 24 | 1.09 (0.49 2.47) | 0.83 |
| Per additional 1 SD (trend) | 0.96 (0.89 1.04) | 0.33 | 0.98 (0.87 1.11) | 0.74 | 0.96 (0.78 1.18) | 0.68 | |||
AGA, appropriate for gestational age; CI, confidence interval; HR, hazard ratio; LGA, large for gestational age; SD, standard deviation; SGA, small for gestational age.
Adjusted for shared familial (genetic and/or environmental) factors, in addition to birth year, sex, gestational age at birth, fetal growth, birth order, maternal and paternal age and education and family history of suicide.
Secondary analyses
In the entire cohort, no interaction was found between preterm birth and SGA in relation to suicide risk on the additive (P = 0.62) nor on the multiplicative (P = 0.64) scale (Supplementary Table S3, available as Supplementary data at IJE online). The absence of additive interaction suggests that SGA accounted for a similar number of suicides among persons born preterm compared with full-term.
When confirmed suicides and deaths of undetermined intent were analysed separately, the risk estimates were overall similar but less precise compared with those from the primary analyses of these outcomes combined. For example, comparing SGA vs AGA, the adjusted HR for risk of confirmed suicide was 1.17 (95% CI, 1.05, 1.30; P = 0.003; Supplementary Table S4, available as Supplementary data at IJE online) and for risk of death of undetermined intent was 1.20 (95% CI, 1.01, 1.44; P = 0.04; Supplementary Table S5, available as Supplementary data at IJE online). Among females, the adjusted HR point estimate comparing preterm vs full-term birth remained elevated whether focusing on confirmed suicides (1.29; 95% CI, 0.98, 1.71; P = 0.07; Supplementary Table S4) or deaths of undetermined intent (1.51; 0.91., 2.48; P = 0.11; Supplementary Table S5). These findings suggest that combining these outcomes in the primary analyses did not strongly influence the results and conclusions.
Discussion
In this large national cohort, preterm birth was associated with an increased risk of suicide death in adulthood among women (but not men) after adjusting for other perinatal and sociodemographic factors and family history of suicide. Low fetal growth was associated with a modestly increased risk of suicide death among both men and women after adjusting for the same covariates. However, these associations were markedly attenuated and no longer significant in co-sibling analyses that controlled for shared familial (genetic and/or environmental) factors. These findings suggest that the associations observed in the entire cohort are not causal, but are related to unmeasured familial exposures that are shared determinants of preterm birth or low fetal growth and suicide risk later in life.
To our knowledge, this is the largest study of these perinatal factors in relation to suicide in a cohort that was followed up into adulthood, and one of the first to explore sex-specific differences and causality. A previous Swedish study examined the association between gestational age at birth and suicide attempts among persons born in 1973–2008 who were followed up through 2009 (N = 3.3 million, of whom approximately half reached adult ages).15 That study reported a nearly 2-fold risk of suicide attempt among those born extremely preterm (23–27 weeks) compared with those born at 37–42 weeks (unadjusted HR, 1.7; 95% CI, 1.2, 2.4), which was completely attenuated in a co-sibling analysis.15 In a subsequent study of the same cohort, low birthweight (<2500 g) also was associated with a modestly increased risk of suicide attempt before but not after adjusting for shared factors among siblings.16 However, those studies did not assess potential sex-specific differences nor suicide death. A recent meta-analysis9 reported pooled unadjusted odds ratios of 1.11 (95% CI, 0.98, 1.25) for suicide death associated with preterm birth (based on five studies17–21) and 1.18 (95% CI, 1.00, 1.40) for suicide death associated with SGA (based on four studies17–20). However, adjustment for confounders varied so widely among studies that pooled adjusted risks were not estimable. Other studies22,23 but not all24,25 have reported associations between low birthweight (<2500 g) and increased suicide risk in adolescence or adulthood, with a pooled odds ratio of 1.30 (95% CI, 1.09, 1.55).9
The present study extends earlier evidence by examining the independent effects of gestational age and fetal growth in a cohort of 2.4 million adults with over 4000 deaths by suicide. The large size of this cohort enabled well-powered analyses of sex-specific differences and adjustment for shared exposures among siblings. The associations that we found between low fetal growth and suicide death are overall consistent with those previously observed in smaller cohorts9 or in relation to suicide attempts.16 In the present cohort, the magnitude of this association was similar in men and women. In addition, we found that women (but not men) who were born preterm had an increased risk of suicide death, which was driven by a 2-fold risk among those born early preterm (22–33 weeks). To our knowledge, this is a new finding but it has rarely been examined in large cohorts of women, and thus will need confirmation in future studies.
Shared familial exposures that potentially may account for the positive associations that we observed include maternal psychosocial stress, other prenatal environmental exposures and possibly genetic factors.26 Maternal psychosocial stress has previously been associated with low offspring birthweight and preterm birth.27 These birth outcomes also have been linked with persistent alterations of hypothalamic-pituitary-adrenal (HPA) axis function in the offspring,28,29 which may contribute to an increased risk of depression and suicidality.30–33 Maternal smoking is associated with low fetal growth34 and to some extent with preterm birth, 35 although some evidence has suggested it has little effect on future suicide risk in the offspring.17 Other maternal substance use36 and gene-environment interactions26,37,38 also may affect birth outcomes and their behavioural sequelae later in life. Clinical and epidemiological studies with additional information on psychosocial stress and other prenatal exposures are needed to identify risk factors shared by siblings which account for our findings. Suicide risk models have seldom included perinatal risk factors.9 Further elucidation of such factors could potentially contribute to a life course approach in suicide risk assessment and facilitate earlier preventive actions in high-risk subgroups.
A key strength of the present study was its large national cohort design, which afforded high statistical power needed to examine narrowly defined exposure groups and sex-specific differences. Highly complete nationwide birth and death data helped minimize potential selection or ascertainment biases, enabling more robust estimates based on a national population. We were able to control for multiple covariates and assess the influence of unmeasured shared familial factors using co-sibling analyses.
This study also had several limitations. We lacked information on specific maternal psychosocial stressors that potentially may contribute to the associations we observed with suicide risk. Studies with access to this information, which include follow-up into adulthood, are a high priority for future research. As in other large population-based studies, the reporting of suicides involves some misclassification. However, available data on intentional deaths and deaths of undetermined intent enabled separate analyses of these outcomes in a sensitivity analysis, which showed little difference in risk estimates. This study also was limited to Sweden and will need replication in other large diverse populations when feasible.
In summary, in this large national cohort, preterm-born females and both males and females with low fetal growth had increased risks of suicide in adulthood. These associations were independent of other measured perinatal and sociodemographic factors, but appeared to be explained by unmeasured shared genetic or prenatal environmental factors within families. Additional studies are needed to identify common modifiable exposures that contribute to preterm birth or low fetal growth and future risk of suicide.
Supplementary data
Supplementary data are available at IJE online.
Funding
This work was supported by the National Institute on Alcohol Abuse and Alcoholism [R01 AA027522 to A.E. and K.S.] and the National Heart, Lung, and Blood Institute [R01 HL139536 to C.C. and K.S.] at the National Institutes of Health; the Swedish Research Council; the Swedish Heart-Lung foundation; and ALF project grant, Region Skåne/Lund University, Sweden. The funding agencies had no role in the design and conduct of the study; in the collection, analysis and interpretation of the data; nor in the preparation, review or approval of the manuscript.
Data availability
Due to legal concerns, the supporting data (which come from a large portion of the Swedish population) cannot be made openly available. Further information about the data registers is available from the Swedish National Board of Health and Welfare [https://www.socialstyrelsen.se/en/statistics-and-data/registers/].
Author contributions
All authors contributed to the study concept and design, interpretation of data, critical revision of the manuscript for important intellectual content and obtaining funding, and approved the final version to be published. J.S. and K.S. also acquired the data and contributed to the statistical analysis, and C.C. drafted the manuscript.
Conflict of interest: None declared.
Supplementary Material
References
- 1.World Health Organization. Suicide Prevention. 2020. [cited June 22, 2020]; https://www.who.int/health-topics/suicide (22 June 2020, date last accessed).
- 2. Turecki G, Brent DA.. Suicide and suicidal behaviour. Lancet 2016;387:1227–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Harris EC, Barraclough B.. Suicide as an outcome for mental disorders. A meta-analysis. Br J Psychiatry 1997;170:205–28. [DOI] [PubMed] [Google Scholar]
- 4. Crump C, Sundquist K, Sundquist J, Winkleby MA.. Sociodemographic, psychiatric and somatic risk factors for suicide: a Swedish national cohort study. Psychol Med 2014;44:279–89. [DOI] [PubMed] [Google Scholar]
- 5. Robinson R, Lahti-Pulkkinen M, Schnitzlein D. et al. Mental health outcomes of adults born very preterm or with very low birthweight: A systematic review. Semin Fetal Neonatal Med 2020;25:101113. [DOI] [PubMed] [Google Scholar]
- 6. Crump C, Winkleby MA, Sundquist K, Sundquist J.. Preterm birth and psychiatric medication prescription in young adulthood: a Swedish national cohort study. Int J Epidemiol 2010;39:1522–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Crump C, Sundquist J, Winkleby MA, Sundquist K.. Gestational age at birth and mortality from infancy into mid-adulthood: a national cohort study. Lancet Child Adolesc Health 2019;3:408–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Barker DJP. Mothers, Babies and Health in Later Life. 2nd edn. London: Churchill Livingstone, 1998. [Google Scholar]
- 9. Orri M, Gunnell D, Richard-Devantoy S. et al. In-utero and perinatal influences on suicide risk: a systematic review and meta-analysis. Lancet Psychiatry 2019;6:477–92. [DOI] [PubMed] [Google Scholar]
- 10.Statistics Sweden. The Swedish Medical Birth Register. https://www.socialstyrelsen.se/en/statistics-and-data/registers/register-information/the-swedish-medical-birth-register/ (4 June 2020, date last accessed).
- 11. Stock S, Norman J.. Preterm and term labour in multiple pregnancies. Semin Fetal Neonatal Med 2010;15:336–41. [DOI] [PubMed] [Google Scholar]
- 12. Ohberg A, Lonnqvist J.. Suicides hidden among undetermined deaths. Acta Psychiatr Scand 1998;98:214–18. [DOI] [PubMed] [Google Scholar]
- 13. Bjorkenstam C, Johansson LA, Nordstrom P. et al. Suicide or undetermined intent? A register-based study of signs of misclassification. Popul Health Metr 2014;12:11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. VanderWeele TJ. Causal interactions in the proportional hazards model. Epidemiology 2011;22:713–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. D’Onofrio BM, Class QA, Rickert ME, Larsson H, Långström N, Lichtenstein P.. Preterm birth and mortality and morbidity: a population-based quasi-experimental study. JAMA Psychiatry 2013;70:1231–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Class QA, Rickert ME, Lichtenstein P, D'Onofrio BM.. Birthweight, physical morbidity, and mortality: a population-based sibling-comparison study. Am J Epidemiol 2014;179:550–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Cnattingius S, Svensson T, Granath F, Iliadou A.. Maternal smoking during pregnancy and risks of suicidal acts in young offspring. Eur J Epidemiol 2011;26:485–92. [DOI] [PubMed] [Google Scholar]
- 18. Danziger PD, Silverwood R, Koupil I.. Fetal growth, early life circumstances, and risk of suicide in late adulthood. Eur J Epidemiol 2011;26:571–81. [DOI] [PubMed] [Google Scholar]
- 19. Lahti M, Eriksson JG, Heinonen K. et al. Late preterm birth, post-term birth, and abnormal fetal growth as risk factors for severe mental disorders from early to late adulthood. Psychol Med 2015;45:985–99. [DOI] [PubMed] [Google Scholar]
- 20. Niederkrotenthaler T, Rasmussen F, Mittendorfer-Rutz E.. Perinatal conditions and parental age at birth as risk markers for subsequent suicide attempt and suicide: a population based case-control study. Eur J Epidemiol 2012;27:729–38. [DOI] [PubMed] [Google Scholar]
- 21. Risnes KR, Pape K, Bjorngaard JH, Moster D, Bracken MB, Romundstad PR.. Premature adult death in individuals born preterm: a sibling comparison in a prospective nationwide follow-up study. PLoS One 2016;11:e0165051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Mittendorfer-Rutz E, Rasmussen F, Wasserman D.. Restricted fetal growth and adverse maternal psychosocial and socioeconomic conditions as risk factors for suicidal behaviour of offspring: a cohort study. Lancet 2004;364:1135–40. [DOI] [PubMed] [Google Scholar]
- 23. Riordan DV, Selvaraj S, Stark C, Gilbert JS.. Perinatal circumstances and risk of offspring suicide. Birth cohort study. Br J Psychiatry 2006; 189:502–07. [DOI] [PubMed] [Google Scholar]
- 24. Osler M, Nybo Andersen AM, Nordentoft M.. Impaired childhood development and suicidal behaviour in a cohort of Danish men born in 1953. J Epidemiol Community Health 2008;62:23–28. [DOI] [PubMed] [Google Scholar]
- 25. Li CI, Daling JR, Emanuel I.. Birthweight and risk of overall and cause-specific childhood mortality. Paediatr Perinat Epidemiol 2003;17:164–70. [DOI] [PubMed] [Google Scholar]
- 26. Schlotz W, Phillips DI.. Fetal origins of mental health: evidence and mechanisms. Brain Behav Immun 2009;23:905–16. [DOI] [PubMed] [Google Scholar]
- 27. Koubovec D, Geerts L, Odendaal HJ, Stein DJ, Vythilingum B.. Effects of psychologic stress on fetal development and pregnancy outcome. Curr Psychiatry Rep 2005;7:274–80. [DOI] [PubMed] [Google Scholar]
- 28. Phillips DI, Walker BR, Reynolds RM. et al. Low birthweight predicts elevated plasma cortisol concentrations in adults from 3 populations. Hypertension 2000;35:1301–06. [DOI] [PubMed] [Google Scholar]
- 29. Kajantie E, Feldt K, Raikkonen K. et al. Body size at birth predicts hypothalamic-pituitary-adrenal axis response to psychosocial stress at age 60 to 70 years. J Clin Endocrinol Metab 2007;92:4094–100. [DOI] [PubMed] [Google Scholar]
- 30. Nemeroff CB, Widerlov E, Bissette G. et al. Elevated concentrations of CSF corticotropin-releasing factor-like immunoreactivity in depressed patients. Science 1984;226:1342–44. [DOI] [PubMed] [Google Scholar]
- 31. Jokinen J, Nordstrom P.. HPA axis hyperactivity and attempted suicide in young adult mood disorder inpatients. J Affect Disord 2009;116:117–20. [DOI] [PubMed] [Google Scholar]
- 32. Jokinen J, Nordstrom P.. HPA axis hyperactivity as suicide predictor in elderly mood disorder inpatients. Psychoneuroendocrinology 2008;33:1387–93. [DOI] [PubMed] [Google Scholar]
- 33. Weinstock M. The potential influence of maternal stress hormones on development and mental health of the offspring. Brain Behav Immun 2005;19:296–308. [DOI] [PubMed] [Google Scholar]
- 34. Lobel M, Cannella DL, Graham JE, DeVincent C, Schneider J, Meyer BA.. Pregnancy-specific stress, prenatal health behaviours, and birth outcomes. Health Psychol 2008;27:604–15. [DOI] [PubMed] [Google Scholar]
- 35. Kyrklund-Blomberg NB, Cnattingius S.. Preterm birth and maternal smoking: risks related to gestational age and onset of delivery. Am J Obstet Gynecol 1998;179:1051–55. [DOI] [PubMed] [Google Scholar]
- 36. Shankaran S, Lester BM, Das A. et al. Impact of maternal substance use during pregnancy on childhood outcome. Semin Fetal Neonatal Med 2007;12:143–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Lunde A, Melve KK, Gjessing HK, Skjaerven R, Irgens LM.. Genetic and environmental influences on birthweight, birth length, head circumference, and gestational age by use of population-based parent-offspring data. Am J Epidemiol 2007;165:734–41. [DOI] [PubMed] [Google Scholar]
- 38. Wichers MC, Purcell S, Danckaerts M. et al. Prenatal life and post-natal psychopathology: evidence for negative gene-birthweight interaction. Psychol Med 2002;32:1165–74. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Due to legal concerns, the supporting data (which come from a large portion of the Swedish population) cannot be made openly available. Further information about the data registers is available from the Swedish National Board of Health and Welfare [https://www.socialstyrelsen.se/en/statistics-and-data/registers/].