"Children are more likely to use cannabis if their mother smoked while pregnant," the Mail Online reports.
New research suggests smoking in pregnancy could affect the genes of the child, increasing their risk of substance abuse in later life.
The research focused on a branch of genetics known as epigenetics. Epigenetics is the study of how external factors can affect gene expression – whether genes are switched on and off, and how that impacts genetic activity.
Researchers found epigenetic variations at birth may predict a child's risk of using alcohol, cannabis and tobacco as teenagers. In turn, they say, these epigenetic variations may be caused by smoking in pregnancy.
The researchers say this may explain how environmental factors, such as a mother smoking in pregnancy, could affect her child's later behaviour.
The research was based on 244 children born in the UK in 1991 or 1992 taking part in a long-term study of health and environment.
But these findings come with a number of cautions. The researchers say they can't prove smoking caused the gene expression variations, or that these variations cause cannabis, alcohol or tobacco use in later life.
Rather, they say their research shows a potential epigenetic pathway that needs further investigation.
We already know smoking in pregnancy is bad for the baby and can cause many health problems, including a higher risk of miscarriage, stillbirth and developmental problems.
If you're pregnant or trying for a baby, the best thing you can do for your baby's health is to stop smoking.
Read more about stopping smoking during pregnancy.
The study was carried out by researchers from King's College London, Exeter University, University College London and the University of Bristol in the UK, and Georgia State University in the US.
It was funded by the National Institute of Child and Human Development.
The reporting by the Mail Online is mostly accurate and balanced, despite a headline that makes it sound as if the study established cause and effect between smoking in pregnancy and cannabis use in adolescence.
But the news story does include the researchers' warning that the study cannot show cause and effect.
The headline also only focused on the illegal drug cannabis without mentioning there were also links with alcohol and tobacco, both of which are also harmful.
This prospective cohort study followed a group of children from before birth until adulthood. Cohort studies are good ways to investigate potential links between factors.
However, it's difficult to prove that one factor causes another as there are often many confounding factors that could provide alternative explanations for the results.
Researchers used information about children whose mothers had joined a long-term study before they were born.
The information was taken at various time points, including samples of the child's DNA taken at birth and age seven, and teenagers' own reports of whether they'd used alcohol, cigarettes or cannabis at age 14, 16 and 18.
The researchers also looked at factors affecting the mother before birth, including smoking, mental health and stressful events.
They looked for links between DNA activity, teenagers' use of drugs or alcohol, and factors affecting the mother.
The children were all part of the Avon Longitudinal Study of Parents and Children (ALSPAC), which began in the early 1990s in the Bristol area in the UK.
Children born in 1991 or 1992 in that area whose parents agreed are all tracked with various measures.
Although the original group included 14,000 children, only 244 had all of the measures required to be included in this study.
The DNA analysis looked for specific areas of DNA called gene loci that had more or less methylation activity than normal, described as epigenetic variation.
Methylation is the attachment of a methyl molecule to gene loci, and is part of the control system that switches genes on or off in cells. This decides how the cell grows and behaves.
Researchers compared patterns of epigenetic variation at birth in blood from the umbilical cord with patterns in blood taken at age seven.
They looked at the functions of genes affected and at possible mechanisms for this epigenetic variation during pregnancy.
The researchers found epigenetic variation in certain networks of genes found in a baby's blood at birth were linked to a higher chance of using tobacco, cannabis or alcohol in teenage years.
However, they did not find these same epigenetic variations in blood taken from the children at age seven.
Children whose DNA at birth showed epigenetic variation were also more likely to start using substances at an earlier age.
The researchers say the genes most affected at birth included those important for nerve growth, signalling and development, especially in parts of the brain linked to drug-seeking behaviour and addiction.
When they looked at the most common epigenetic variations, they found some of them were linked to a range of factors seen in pregnant women.
The strongest link they found was between higher DNA methylation and smoking during pregnancy.
The researchers say they showed how epigenetic variation at birth was linked to higher tobacco, cannabis and alcohol use in teenagers.
They say these effects were specific to the children's DNA methylation at birth, not later in childhood.
Together, they say, the epigenetic variation in the affected genes "mediated the effect of prenatal tobacco smoking on adolescent substance use".
In other words, the epigenetic changes could be what linked the mothers' smoking and the child's use of drugs and alcohol.
They caution that "such evidence should be considered preliminary", the results "should be interpreted with caution", and "it is not possible to establish causality" from their research.
Research into epigenetics helps scientists investigate how the genes we inherit from our parents interact with the environment around us – for example, how genes combine to make us more or less likely to behave in certain ways or develop certain conditions.
We know from experience that some people seem more susceptible to risks like becoming addicted to alcohol or using drugs or tobacco.
This research may help us understand some of the factors behind that difference in risk, although it is likely there are many different causes.
Peer pressure, social and economic circumstances, parental attitudes, laws and the price of substances are all likely to affect whether or when a teenager starts to use tobacco, cannabis or alcohol. Still, epigenetic variation may be one more factor to consider.
This study doesn't tell us for sure whether smoking in pregnancy causes epigenetic changes that put teenagers at risk of substance abuse.
However, we already know smoking in pregnancy has a wide range of harmful effects on the health of the unborn baby by restricting oxygen and introducing toxins into the growing child.
And it can also increase the chance of developmental problems, stillbirth and miscarriage.
If you're pregnant or want to get pregnant and you smoke, the best thing you can do for your baby is to stop smoking.
There's lots of help available. Ask your doctor or midwife, or find out about NHS Stop Smoking services.