The intergenerational effects of war on the health of children

Violence is ubiquitous in war and pervades all strata of society, although its occurrence is probably underestimated [20‐22]. The 20th century saw a secular trend toward a greater proportion of civilians affected, especially women and children [23]. It is estimated that up to 90% of war-related deaths in the last decade were civilian [5]. Conflict creates an ecosystem that persists, but is maladapted to peacetime, and its legacy is frequently an environment that fosters violence. Numbers of small arms remain for years, an example being the 1996 civil war in Guatemala, where rates of gun violence were greater after the war ended [24]. Similarly, increasing armed conflict in land disputes in East Africa is a consequence of prior cold war politics [25], and the 110 million landmines currently thought to remain in 68 countries [26] may lie dormant and harm civilians, both physically and mentally, decades after the original conflict [27, 28].

Women are at particular risk of combatant and civilian injury, and of interpersonal violence. Violence against women increases in times of conflict through targeted acts, such as rape and domestic or intimate-partner violence. Sexual violence, including rape, assault, trafficking and prostitution, increases during many conflicts due to the breakdown of traditional safety structures [29]. It may also be used as an intentional strategy of domination, as reported in the former Yugoslavia and Rwanda [20, 30]. The magnitude of sexual violence in conflict zones is difficult to quantify and tends to be under-reported [31]. Approximately half of a sample of women interviewed in Liberia had suffered sexual violence [21] and studies in East Timor and Kosovo showed increases to 23% and 15%, respectively, during conflict, compared to 10% and 2% in the post-conflict period [32, 33]. A 2006 study from Lebanon suggested that domestic violence increased in areas where there was collective violence [22]. There are other, more complex, links. Substance use is associated with violence against women in non-conflict settings [34‐36], although the evidence that it increases during conflict is limited. Clearer is the bidirectional association between violence and alcohol use [37‐39].

A clear example of an intergenerational effect is stillbirth after physical violence against a pregnant woman [40, 41]. During the siege of Sarajevo, perinatal mortality and morbidity more than doubled, and there was a rise in congenital malformations from 0.4% to 3% [42]. Violence is also associated with an increase in premature birth [43], as was seen during the war in Croatia in the early 1990s [44]. Similar to this in non-conflict settings, a meta-analysis on abuse during pregnancy showed an increased odds ratio for low birth weight of 1.4 (95% confidence interval (CI) 1.1 to 1.8) [45].

Sexual violence has both short- and long-term effects. In the short-term, it is associated with an increase in sexually transmitted diseases, the intergenerational effects of which are discussed later. Maman and colleagues relate HIV infection to violence through three mechanisms: ‘forced or coercive sexual intercourse with an infected partner; by limiting women’s ability to negotiate safe sexual behaviors; and finally by establishing a pattern of sexual risk taking among individuals assaulted in childhood and adolescence’ [46]. Even when services are available, fear of violence may lead to a lower likelihood of accessing treatments to prevent mother-to-child transmission of HIV [47]. In the longer-term, sexual violence may affect parenting capacity. In Rwanda and Darfur, children born after rape have been shown to be at increased risk of neglect, abuse, malnutrition and abandonment [48]. Evidence from hospitals in South Kivu, Democratic Republic of Congo, suggests that when children are born from rape, their mothers are five times more likely to suffer isolation from the community (odds ratio (OR) 4.84; 95% CI 1.41 to 19.53) [49]. Children may also be victims of sexual violence [9], which may change its characteristics in times of conflict: a greater likelihood of gang rape was seen, for example, during conflict in the Democratic Republic of Congo. Sexual violence against children has profound effects on their mental and physical development, with implications for future pregnancies and relationships [50]. Mental illness, affecting a family’s ability to function, can persist long after conflict has ended: symptom rates in those who had suffered from mass violence remained elevated for a decade after the Cambodian conflict [51]. More generally, exposure to violence has been associated with long-term effects on children’s health, brain structure and neural function [52]. A violent environment may also have long-term effects on the maturational schedule. In Colombia, the secular decline in age at menarche accelerated during periods with high homicide rates, suggesting that psychosocial cues can alter the developmental trajectory [53].

Violence can affect family life, both economically and socially. Among Cambodian families in which one member was injured by a landmine, 61% were pushed into debt [54]. A study from Nicaragua found that women who suffered from abuse had 44% lower earnings [55]. Combat, death and displacement can also lead to the breakdown of family structures. Family members may shift roles, as was noted among women in Nepal who became combatants and replaced those lost during the conflict [56].

It is generally thought that experiencing violence as a child is a risk factor for committing child abuse as an adult [57]. There is some evidence that children who witnessed frightening events (for example, in the Spanish civil war of 1936) are ‘more immune to the horrors of violence’ [4]. This is especially so for child soldiers, who may experience dehumanizing conditions during a formative period, but there is a danger that maladaptive behaviors are passed on to the next generation. There are few long-term studies in this group. In a 16-year follow-up of child soldiers in Mozambique, men still suffered distress from their traumatic events, but there was no evidence that abusive relationships were passed on to their children [58].

Chemical or radioactive weapons can also damage future generations. While evidence of genetic effects of radiation on the children of survivors of atomic bombs is lacking, follow up studies have found evidence of severe learning difficulties and microcephaly in people exposed in utero. In utero exposure was also associated with decreased school performance and intelligence quotient (IQ) [59, 60]. Case–control studies suggest that exposure of men to mustard gas in Iran and mineral contamination (particularly depleted uranium) in Iraq was associated with congenital abnormalities in their children [61, 62].

Mental health conditions are common during conflicts, and are augmented by the breakdown of mental health services and community coping strategies, and by increases in stress levels and drug and alcohol use [63]. The most common conditions are depression, anxiety and psychosomatic disorders, but the most widely studied in relation to conflict is post-traumatic stress disorder (PTSD). PTSD is common among combatants, and child soldiers are known to suffer from a much higher than average incidence of psychological disturbances and mental illness [64‐67]. PTSD is also found in civilian populations in both women and children [68, 69]. De Jong et al. showed that PTSD and anxiety disorder were the most common mental illnesses in all populations, and exposure to violence increased their likelihood in the four conflict zones of Algeria, Cambodia, Ethiopia and Palestine [70]. In Kabul, mothers of children under five were found to have an increase in prevalence of PTSD from 10% to 53% if they experienced at least one armed conflict event [71], and women living in conflict areas in Bosnia and Herzegovina showed a sevenfold increase over a control group in PTSD prevalence [72]. Importantly, mental health disorders also rise in those not directly exposed to violence. Women are generally affected more than men and strong associations exist between the mental health of mothers and their children [63]. For example, in Algeria poor quality of camp housing and the general difficulties of daily living were both associated with increased odds of PTSD (OR 1.8, 95% CI 1.3 to 2.5 and 1.6, 95% CI 1.1 to 2.4, respectively) [73]. Mental illness is also common in children. A systematic review in a long-term conflict in the Middle East showed that children do suffer from a wide range of mental disorders, such as PTSD, depression, anxiety, functional impairment, behavioral problems and Attention Deficit Hyperactivity Disorder [74].

The evidence on drug use in conflict is partial. Drug use increases in groups that have been displaced, especially among combatants and former combatants, for which it may be a way of coping with traumatic situations or memories [75‐78]. Economic pressures and increased availability can lead to a rise or change in drug use. For example, the conflict and related drug policy in Afghanistan have been linked to an increase in intravenous opiate use [79, 80]. The pattern of illicit drug use depends on the setting: opiates are more common in Pakistan and Afghanistan, while benzodiazepines are common in Bosnia-Herzegovina [81]. A study of khat (a substance that contains a psychoactive compound similar to amphetamines) in north-west Somalia showed that recent use among ex-combatants was 60%, compared with 28% in civilian war survivors and 18% in civilians with no experience of war. Psychotic symptoms were linked with age of onset of khat use and binge consumption. Rates of severe disability due to mental disorders (primarily psychosis) in males over the age of 12 increased in relation to their experience of war, with a prevalence of 16% in ex-combatants, 8% in civilian war survivors and only 3% in those with no experience of war [82].

While evidence is currently limited, parental trauma and psychosocial stress during conflict have been associated with adverse health effects in offspring, through both biological pathways (such as neuroendocrine and immune system modulation) and propagation of stressful social environments [83‐85]. Stress may be a mediator through which parents transmit adverse effects to their children, in both conflict and non-conflict settings. In a large population-based cohort study, maternal stress during pregnancy was associated with a number of outcomes in children, including risk of infection and mental disorders [86]. Stress arising from childhood separation from parents during World War II was linked to long-term impairment in offspring social mobility and socio-economic position [87]. A study in people living near Gaza showed a strong association between maternal symptoms of depression, anxiety and PTSD, and PTSD symptoms in children. Avoidance behaviors in children were also associated with the degree of trauma exposure in the mother [88], similar to associations shown in animal studies [89]. Possible mechanisms include epigenetic changes in the hypothalamic-pituitary-adrenal axis (see Box 1) [90, 91]. There is also some evidence that paternal PTSD can likewise lead to symptoms in children [92].

Stressful emotions can be passed from one generation to the next, but maternal buffering appears to provide an important damping effect, so that children are less affected than their parents. Children exhibit a form of resilience [93] that is related in turn to their mothers’ wellbeing, quality of caregiving and environmental support [94], but traumatic events, such as those occurring in a conflict, may override this buffering. For example, children of Australian veterans of the Vietnam war were found to have a suicide rate three times that of the general community [95] and elevated rates of PTSD and other mental illness have been described in the offspring of Holocaust survivors [96‐100]. Mothers, but not fathers, who were themselves offspring of Holocaust survivors had ‘higher levels of psychological stress and less positive parenting’, while in the grand-offspring generation, children had less positive self-perception and were said by their peers to show ‘inferior emotional, instrumental, and social functioning’ [101].

There is also evidence of intergenerational mental health effects in non-conflict settings, and these are likely to be exacerbated by conflict. A meta-analysis by Surkan et al. showed that maternal depression or depressive symptoms were associated with an OR of 1.5 (95% CI 1.2 to 1.8) for children being underweight and 1.4 (95% CI 1.2 to 1.7) for stunting. The reasons for this include poorer antenatal care, increased risk-taking behaviors and impaired maternal caring [102], with reduced or early cessation of breastfeeding as one possible mediator [103, 104]. Maternal mental illness has also been linked to children’s cognitive abilities [105]. In India, maternal common mental disorder was associated negatively with mental development in infants at six months of age [106] and, in Barbados, with cognitive outcomes [107]. The latter association persisted when assessed via a school entrance exam at 11 to 12 years of age [108].

There are several mechanisms whereby drug use may exert adverse intergenerational effects. Drug use by pregnant women can have transplacental effects or cause maternal ill-health or altered behavior. The manifestations may be acute - neonatal abstinence syndrome from opiate withdrawal, for example [109] – or lead to longer-term behavioral and cognitive changes [110]. Both drugs and alcohol are closely associated with mental illness in the user, which, in turn, can have detrimental effects on parenting ability and employment. The use of khat has been linked to mental illness, as well as affecting the reproductive system and being teratogenic [111]. However, the magnitude of the intergenerational effect of drug use associated with conflict remains to be established.

Mortality in women and children during conflicts is predominantly the result of conditions related indirectly to violence: infectious disease, malnutrition and complications of pregnancy [8]. With the exception of a few examples for which limitation of movement has appeared to prevent transmission [112], higher rates of infectious diseases, such as shigellosis and cholera, are common in conflict zones [113, 114]. As described in Table 1, the major drivers are movement of people, infrastructural deterioration, and breakdown in prevention and treatment [115]. Migration of people, for example, both increases infectious disease burden and exposes unimmunized populations to new pathogens.

Sexual violence against women has been mentioned, and sexually transmitted infection rates may rise in conflict settings, as seen with syphilis and gonorrhea in the Second World War [129]. Increased infections are especially common in refugee or IDP populations [130], and in situations characterized by alcohol use and riskier sexual practices [75]. They have also been linked with psychiatric conditions and psychosocial factors. In war-affected Eastern Uganda, major depressive disorder and sexual torture were associated with high-risk sexual activities, which were themselves associated with HIV transmission [131]. It is thought that the conflict in Mozambique led to higher rates of syphilis in pregnant women, some of whom experienced sexual abuse and repeated rape while being held by the insurgents [132]. HIV rates may also be augmented by untreated sexually transmitted infections, lack of condoms, increased incidence of sex work and lack of health education [80, 119]. An estimated 70% of Rwanda’s rape survivors were infected with HIV [133].

Infectious diseases can affect subsequent generations through direct effects on pregnant women or through the effects of long-term morbidity on their future health, reproductive capacity and finances. The Spanish influenza outbreak, which resulted in up to 40 million deaths, is considered to have originated in and been exacerbated by the conditions of the First World War [134, 135], and research on prenatal exposure suggests both long-term health costs, such as increased rates of cardiovascular disease and potentially increased disability, and economic penalties [136, 137]. More specifically, infection in utero may lead to abortion or congenital malformations, and many infections can trigger premature birth. A fall in immunization rates may be accompanied by reductions in herd immunity: outbreaks of rubella, which can lead to severe congenital abnormalities, occurred during the conflict in Bosnia and Herzegovina [138]. Congenital infections lead to impaired cognitive development, respiratory and gastrointestinal disease, and may leave children vulnerable to infections during their lifetime [139].

Infections such as HIV and syphilis can be transmitted vertically from mother to child, causing acute infections, fetal or infant death or chronic childhood conditions. A review of neurodevelopmental outcomes showed that vertical transmission of HIV was associated with poorer neurodevelopmental outcomes in children from resource-poor countries [140]. Likewise, repeated malarial infections and continuing parasitemia are associated with reduced cognitive ability in children in the short- and long-term, with effects on their future economic capacity [141‐143].

Lack of human resources adversely affects any future health system. Conflict-driven collapse of the education system, for example, reduces the pool of people who could become health workers. The best-known example is the breakdown of the education system under Cambodia’s Khmer Rouge regime, in which teachers were systematically imprisoned or executed. This not only affected the education of children at the time, but also led to a longer-term loss of knowledge and skills from society.

As common users of healthcare services, pregnant women are especially vulnerable to healthcare service disruptions, leading to the exacerbation of illness. An example of this would be the disruption to the supply of drugs and other medical equipment, as described in northern Sri Lanka during the recent civil war. Obstetric complications, such as pre-eclampsia, were common, potentially contributing to the higher rate of low birthweight that was found [144]. In Palestine, military checkpoints delaying access to maternity facilities were associated with an increase in home births from 8% in 1999 to 33% in 2002 [145], while in Nepal’s recent civil war, disruption of a hospital’s electricity supply led to Caesarean sections being performed by torchlight [123]. More generally, maternal mortality rises during conflict. About half the countries suffering from conflict have maternal mortality ratios >200 per 100,000 live births [8]. The maternal mortality ratio during the conflict in Bosnia and Herzegovina was approximately 75% higher than in the periods immediately before and after [146]. A catastrophic event in its own right, maternal mortality also has damaging effects on the remaining family members, particularly children. Evidence from a non-conflict area of Bangladesh shows very high mortality among younger children - up to 70% for a surviving neonate within two years of the maternal death - approximately 10 times higher than if neither parent had died. Higher mortality persists in children five- to ten-years old, for whom the probability increases five-fold [147]. Similar evidence from Haiti shows increased odds of mortality after a maternal death of 55% for children younger than 12-years-old [148].

Malnutrition is common in conflicts and is predominantly manifest in children [149]. During a recent famine in the Horn of Africa, rates of global and severe acute malnutrition reached levels of 50% and 35%, respectively [150]. Food security falls during conflict for reasons ranging from reduced crop production to breaks in the supply chain and trade restrictions [9, 151]. Food resources may be destroyed deliberately as a means of harm or population control: in 1980, 140,000 hectares were destroyed by Ethiopian government forces to prevent their use by rebel groups [152]. The diversion of resources away from healthcare and food supply to military expenditure in war can adversely affect population health. Despite an increase in overall gross development product per capita, the early Nazi regime saw an increase in mortality and a decline in child height [153]. International sanctions, a favored government response to civil conflict, may generate similar effects. Neonatal mortality, for example, increased substantially in Iraq during the 1990s [154].

There is some evidence that breastfeeding rates can decline in war times and the adverse effects of this can be greater. In the war in Croatia in the early 1990s, breast-feeding duration was found to be shorter in affected areas [155]. In Guinea Bissau in the late 1990s, infants who were weaned earlier during war periods had higher mortality rates than breastfed infants compared to the pre-war years where this difference was not present [156]. In addition to not having breast milk, we would postulate that disruption to formula milk supplies and unclean water would compound the adverse effects. It was shown that children who were never breastfed were more likely to be malnourished [157].

War may contribute to either acute malnutrition and increased mortality or chronic malnutrition leading to stunting and subtle prolonged deficits associated with lower school attainment and reduced adult income [158]. Nutritional deficiencies may be in macronutrients, providing the basic energy and substrate for growth, or in micronutrients, including vitamins and minerals that promote cellular function. Micronutrient deficiency diseases are common in conflicts: scurvy in Afghanistan from vitamin C deficiency [159], pellagra in Angola from niacin deficiency [160].

Malnourished mothers pass on the stress to their children, whose poor nutritional status may affect subsequent generations. In the Dutch Hunger Winter of 1944 to 1945 during the German occupation, maternal under-nutrition was associated with increased risk of low birth weight [161]. Although not directly attributed to malnutrition, a higher proportion of babies were born with moderate low birth weight in Croatia during the war period (1991 to 1995) [162]. Low birth weight is the strongest risk factor for mortality in early infancy [163] and is associated with reduced educational attainment and physical work capacity [158]. Subtler effects include maternal micronutrient deficiencies, such as iron deficiency anemia [164], and conditions that directly affect the fetus, such as neural tube defects associated with folate deficiency [165]. Prenatal malnutrition, leading to low birth weight, results in lower subsequent height and lean mass [166]. Short maternal stature is associated with an increased risk of gestational diabetes, macrosomia, and birth injury and shorter gestation [167‐169]. Among Hmong refugees from the Second Indochina war, children displaced during infancy were shorter as adults, whereas children born during the war were found to have greater adiposity, particularly central adiposity [170]. Such growth penalties may take generations to resolve. Maternal short stature is a risk factor for obstructed labor, Caesarean section and low birth weight [171‐173], potentially generating a long-term intergenerational cycle [15]. Contrasting with sperm production in men, women’s ova are formed primarily early in life, and damage to the reproductive system in young girls, from infections, trauma or harmful substances, for example, can have long-term effects on their reproductive capacity and their children [174, 175].

Peri-conceptional, fetal and infant malnutrition can affect the risk of non-communicable disease. People exposed to maternal malnutrition in utero during the Dutch Hunger Winter showed increased risk of obesity, hypertension, cardiovascular disease and type 2 diabetes [176‐178]. The longer-term effects depended on when in gestation the lack of nutrition occurred and the period for which there was a deficit, and there is some evidence for a difference in the long-term effects between male and female offspring [179, 180]. Although the evidence is conflicting, it may be that the detrimental effects of the Dutch famine extended into the following generation, with the offspring of women born in the famine found to have a lower birth weight [161, 181]. Similar associations between early-life famine exposure and subsequent elevated chronic disease risk were reported following the Biafran conflict of 1968 to 1970 [182], and the Chinese famine of 1959 to 1961 [183‐186]. In the latter, early-life exposure was also associated with increased risk of schizophrenia [187, 188]. In non-conflict settings it is common for chronic disease in adults to affect their offspring, both biologically and economically via loss of earnings. An example of this would be the propensity to obesity and the development of gestational diabetes leading to preterm birth and macrosomia.

Deliberate military efforts may not only impair the capacity for food production but also increase exposure to toxins. For example, bombing and Agent Orange were used in Vietnam to ‘deny the enemy sources of food and means of cover’ [189]. Approximately 11 to 12 million gallons of Agent Orange were sprayed over nearly 10% of South Vietnam between 1961 and 1971. While research on intergenerational effects continues, the WHO describes dioxin, the active element of Agent Orange, as ‘highly toxic’ and able to cause ‘reproductive and developmental problems’ [190, 191].