PhD student investigates causes of premature birth

/ Author: Jan Hendriks
Premature twins in the neonatology department of LUMC
© Hollandse Hoogte / Marc de Haan
On 18 January 2022, Eduardo Villamor Martinez obtained his PhD cum laude from Maastricht University. Data scientist Villamor, who works for Statistics Netherlands (CBS), successfully defended his PhD thesis on ‘Endotypes and complications of very and extremely preterm birth’. Using the statistical methods of meta-analysis and meta-regression, he analysed a large number of relevant studies and other research and generated promising insights that could lead to improvements in neonatology.

Complications

‘Babies who are born prematurely have a higher risk of developing complications, both as children and later in life,’ Villamor discovered. ‘They also have a greater chance of dying prematurely.’ Those are the key conclusions from this CBS data scientist’s thesis. To reach those conclusions Villamor examined results from several hundred relevant studies and other research. These studies contained a vast amount of data, which the PhD student incorporated into his analysis using methods including the statistical methodology known as meta-analysis.

Greater risk of health problems

Whereas a normal pregnancy lasts 40 weeks, Villamor based his thesis on data on premature (under 37 weeks), very premature (28-32 weeks) and extremely premature (under 28 weeks) babies. ‘Very premature and extremely premature babies account for just 1.5 percent of the infant population, but 50 percent of child mortality,’ he explains. ‘These two groups of premature babies also have a much higher chance of developing health problems and needing long-term care. The more premature the birth, the greater the risk of health problems, both as children and in later life.’

Causes of premature birth

Villamor discovered that babies born prematurely can be categorised into two endotypes (causes of premature birth). ‘The first cause is an infection of the amniotic fluid, the uterus or the placenta during the pregnancy. The second is a placenta that is not functioning correctly, which is associated with high blood pressure in the mother during the pregnancy.’

Eduardo Villamor Martinez graduated cum laude from UM

Meta-regression

During his research, Villamor examined the six most prevalent complications in premature births: complications related to the eyes, lungs, brain, heart, bowels and infections. ‘The more premature the birth, the greater the chance of developing these complications. That’s why reducing and treating these complications is one of the most important pillars of care for this group of babies.’ Villamor incorporated third factors into his analysis using the statistical method known as meta-regression. ‘For example, if you look at the chance that a baby born prematurely as a result of placenta dysfunction will develop a lung condition, you see that the likelihood decreases in line with how recently the baby was born.’

Lung complications

After taking a closer look at the main differences between the two most common causes of premature birth, the PhD student concluded that, on average, babies who were born prematurely as a result of placenta dysfunction were born half a week later than other premature babies. ‘Despite this later birth, they are actually at higher risk of developing lung complications. When a baby’s growth is delayed, the risk of other complications increases.’ During his research, Villamor revealed that babies born prematurely as a result of an inflammation come into the world an average of 1.2 weeks earlier than other premature babies. ‘This group is at increased risk of the six most prevalent complications.’ He also assessed the effect of potential interventions following a premature birth: ‘For example, milk from donor mothers decreases the risk of lung complications.’

New paradigm

According to Villamor, the most important contribution his thesis makes is in setting a new paradigm in paediatrics for two important endotypes, or causes of premature birth. ‘This research can help predict which clinical problems will arise when babies are born prematurely, and that knowledge should make it possible to personalise treatment more in neonatology wards. Machine learning can play a key role in developing those treatments, and the insights contained in my thesis will help that process moving forwards.’