Famines and stunting: Are adolescents the hardest hit?

The UN recently raised a red flag that we are heading for one of the worst humanitarian crises since 1945. 20 million people in four countries – South Sudan, Somalia, Yemen and Nigeria – face the risk of famine. The world could witness horrors again. Adding to our concern is the prediction that climate change will bring even more weather variability to these regions, interacting with many complex, man-made conflicts and risks.

Children are among the most vulnerable: some will not survive and others will suffer the consequences of undernourishment, including stunting. The severe consequences of neglecting the needs of a child during their first thousand days, is well evidenced. The message about the importance of the first thousand days has echoed widely – perhaps to the point where concerns may be raised about the policy implications being overstretched.

There is an anthropometric impact of famine on those exposed as infants, but surprisingly even more so on those exposed as adolescents.

The world has witnessed famines before. Publicly available household survey data sets, such as UNICEF/MICS and DHS, enable us to shed additional light on the long-term consequences. An article in the American Economic Review (R. Akresh et.al. 2012) presents evidence of the long-term impact of exposure to the Biafra famine (Nigeria 1968-70). The outcome measure is height among adult women, with height loss well known to be a strong predictor of a range of adverse impacts. The study takes into account the regions and ethnicities most exposed to the famine, as well as the age at exposure and its duration.

As expected, those women who were exposed to the famine as infants, are shorter, with the impact estimated to be minus 0.75 cm. Far less expected is that the height impact on those exposed as adolescents – the second period of life marked by a strong physical growth surge – is much greater, an astonishing minus 4.5 cm! The finding should raise some eyebrows, as stunting is almost exclusively discussed as a phenomenon linked to undernourishment among infants; much more rarely as a phenomenon developed in adolescence.

Source: DHS, all available Cambodia data sets (rounds 2000/2005/2010/2014). Lines based on five-year rolling averages by age at exposure, using 1977 for age 0. Minuses if unborn (-5=born 5 years after 1977). All women above 20 at time of survey are included.
Source: DHS, all available Cambodia data sets (rounds 2000/2005/2010/2014). Lines based on five-year rolling averages by age at exposure, using 1977 for age 0. Minuses if unborn (-5=born 5 years after 1977). All women above 20 at time of survey are included.

The graph above conveys a highly simplified visual replication of the Biafra findings, related to another tragic famine: Cambodia 1975-79. In terms of excess mortality its estimated impact is 1.5-2.0 million. In an overview of 20th century famines (S. Devereux 2000) the Cambodia famine stands out in terms of its magnitude and the share of the population that was affected. In a spirit of simplicity, the above graph plots the average height of women (blue line) and the prevalence of adult female stunting (<145 cm, red line) against age in 1977. That year is chosen for being the mid-year of the 1975-79 episode. Data on height and year of birth has been brought together from all available Cambodia DHS surveys. All estimates are national averages; no adjustments have been made of more or less exposed regions or population groups, nor for the length of the exposure to the famine (both adjustments would probably further underscore the message conveyed by this graph). The graph echoes the message from the cited Biafra study: There is, as expected, an anthropometric impact of famine on those exposed as infants, but surprisingly even more so on those exposed as adolescents.

These findings lend themselves to a number of interpretations. A first could be that excess mortality primarily affects infants, which implies that those most severely affected by famine are no longer with us in the data-sets we use. Additional research could probably indicate the extent to which this is a factor at play here. However, notwithstanding its importance, this does not diminish concerns over adolescent vulnerability.

A second interpretation could be related to the much debated issue of catch-up growth. A recent study by Young Lives – based on longitudinal data from Peru, India, Ethiopia and Vietnam – does lend further support to earlier results, which indicate catch-up growth among adolescents who were stunted as infants. So one interpretation may be that we see less height impact on those exposed as infants in the graph above, or in the Biafra study, because catch-up growth has taken place. Furthermore, it may be that catch-up growth does not repair other detrimental impacts of infant stunting, and that the curve somehow masks that. The fast brain development of infants is often cited as a specific cause of concern. However, recent brain science research tends to emphasize the crucial brain growth that also occurs during adolescence. The research on the long-term consequences of stunting depending on at which age it occurs, have important contributions to make here.

In any case, the notion of catch-up growth underscores one unique aspect of adolescents’ vulnerability to famines: Adolescents will not be given a second opportunity to grow.

A third interpretation of the graph above, not necessarily incompatible with the first two, could be that adolescents are indeed disproportionately neglected during famines.  Concerns over such biases are not new (P. Salama 1998). Infants require less food and are possibly given more attention by immediate caregivers, as well as by emergency relief operations, while adolescents tend to be treated as adults, despite their nutrition-related vulnerability. If this is the case, then policies and practices need rethinking.

At UNICEF Innocenti, we were considering setting up our own research project to add more evidence and nuance to the issue discussed here, including replicating it with more precision for other outcome variables and to other incidents of famine. However, given the urgency of the matter, we have chosen to share these incipient findings and puzzles, in a blog. We invite the research community to help us sort out this issue, urgently. Please share your findings and research plans in the comment field below, or contact us directly! If there is a sufficiently strong response, we may consider setting up an Innocenti conference on “Famines and anthropometric impact on adolescents,” hopefully soon enough to make a difference, as the world responds to the upcoming humanitarian disasters.

Goran Holmqvist is Associate Director and Audrey Pereira is a consultant at UNICEF Innocenti. Explore the UNICEF Innocenti research catalogue for new publications. Follow UNICEF Innocenti on Twitter and sign up for e-newsletters on any page of the UNICEF Innocenti website.


Comments are closed.


  1. Important and timely blog.

    On the point “The research on the long-term consequences of stunting depending on at which age it occurs, have important contributions to make here.” A couple of references from work on the Young Lives study of children growing up in Ethiopia, India, Peru and Vietnam are relevant. These papers provide supportive evidence linking post infancy (physical) catch up growth with impacts on later cognitive test performance.
    “The Sooner The Better But It’s Never Too Late: The Impact of Nutrition at Different Periods of Childhood on Cognitive Development”
    “Growth recovery and faltering through early adolescence: Determinants and implications for cognitive development”

    1. Thanks Paul. Longitudinal surveys have a lot to offer on these issues. Just a pity many longitudinal surveys, at least in developing countries, tend to end at adolescence or right after. So we don’t know what happens next. Maybe a few new rounds of Young Lives would be a good idea…

  2. For adolescents there are much wider concerns beyond anthropometry that affect them- learning abilities, self esteem, skills and ability to participate in productive and sporting activities, peer pressure etc. Interventions that have been debated during adolescence e.g. Menstrual hygiene, adolescent school feeding (primary & secondary) are worth more study now.

  3. I think the famine can have a significant impact on adolescent development. During this period the children are transitioning into adulthood as such it is a time of significant physical and social growth. The development and maturation of reproductive organs require sufficient nutrition… weight and height also require steady supply or consumption of food. Famine can therefore compromise adolescent development. We can also discern some social and emotional effects.. famine can result into loss of innocence.. some adolescents can resort to petty theft to get food … The powerful or those with food can sexually take advantage of young girls and boys.

  4. Do you plan to systematize the analysis you have done for Cambodia to all countries where there has been a famine and where DHS or MICS data are available? How do you plan to account for higher mortality rates among under 5 that can skew the results? Do you plan to do time series analysis or to compare to other non affected groups?

    1. Honestly, we have no real plans at this point. The idea with the blog was to raise the issue and find out who else may have explored, or are about to explore, this topic. For sure plenty of things one could do to expand and refine the analysis, including analyzing the excess mortality issue and expanding to other famines an more. Cambodia is a rather clear-cut case simply because of its horrifying magnitudes.

  5. Dear Goran, dear Audrey, your findings are consistent with a previous research by Richard Akresh et al., “War and Stature: Growing Up during the Nigerian Civil War,” American Economic
    Review 102 (2012b): 273–77, doi: 10.1257/aer.102.3.273. They found that 40 years after the war ended, its full consequences were still being realized. Women who had been exposed to the war for
    the average duration between the time they were newborns and 3 years of age were 0.75 centimeters (0.3 inches) shorter than women the same age who hadn’t been exposed. Women who were exposed when they were 13 to 16 years old were 4.53 centimeters shorter. The fact that war exposure in adolescence had the strongest impact is striking. This effect may have stemmed from disruption of the normal adolescent growth spurt. Children’s growth in height is fastest during infancy, slows down until around age 3, and then continues at a low rate until peaking again in adolescence. However, we have limited causal evidence of how nutritional deprivation affects children at different ages, and we particularly lack studies that compare how shocks experienced during adolescence differ from shocks experienced during early childhood. The above is an extract from “Climate change, conflict, and children” by Richard Akresh , 2016″

    1. Yes exactly, the Akresh article is what inspired to write this blog to start with and to test its finding on the Cambodia famine. It is the same American Economic Review article we cite. We discovered it via the 2016 overview article on “Climate Change, Conflict and Children”, which I think provides a good overview of the literature. It points out this gap in understanding broader impact of shocks during adolescence, versus impact of shocks during early childhood, which is key to assess the relevance of these findings on height impact. It should be doable for research to come closer to an understanding of that issue. /Goran Holmqvist

  6. A PhD student here at UNC, Averi Chakrabarti, has been exploring this idea for her dissertation topic so this is a very coincidental and timely blogpost! Currently her idea is to compare the effects of shocks (broadly defined) at time of birth with such shocks at menarchie on adult height among women, to see if the effects are larger when they occur during menarchie. She is exploring data from India and the US for now, maybe other countries later as the work progresses.

    An excellent blog to help us all remember that second (and final) growth spurt that children undergo.

    1. Thanks Ashu. Averi’s work appear to be spot on. We keep the name in mind should we engage in some follow-up work.

  7. Not to minimise the importance of height but long term more debilitating may be the damage to the brain development, this could be particularly significant for children in utero during famine and the first two years and perhaps during the second burst of brain reorganisation during the teens. Is there any information on this. It is clearly much more difficult to study

    1. Agree. Clearly concerns are not about the height per se but what it may predict…and it appears we don’t now enough about how that depends on age at exposure to a shock. Some work is being done on it, but not much as far as we can tell…we are trying to find out if there is more. /Goran Holmqvist