In the last issue of the American Journal of Obstetrics & Gynecology, a large Norwegian study was published by Camilla Haavaldsen MD, et al, which looked at the association of fetal death in relation to maternal age and length of gestation. It is, after all, no secret that in many cultures around the world (not just the U.S.!) women are extending their childbearing years, considerably. According to the March of Dimes, 1 in 5 American women have their first baby at age 35 or beyond. Research then, which assesses pregnancy outcomes in terms of maternal age are important as this trend continues.
The statistics in the Haavaldsen study, not unlike others of similarity (referenced in the publication) are strikingly concerning. The study’s conclusion states,
Women 40 years old or older had the highest risk of fetal death throughout pregnancy, particularly in term and post-term pregnancies.
For the tens of thousands of women who are delaying their childbearing years, the raw data and even relative risk numbers are staggering.
In many ways, this study seems to be well thought out: potentially confounding variables were controlled for, including: parity, plurality, year of delivery, paternal age and existence of pre-eclampsia (defined as maternal blood pressure > 140/90). Following a complex system of data analysis, the blanket results were as follows: of the 2,182,756 pregnancies studied between 1976-2006, 22,754 resulted in fetal deaths—1.049%. Of note, the data came from the Medical Birth Registry of Norway—a compulsory method of tracking perinatal mortality rates which, of course, we here in the U.S. have no federally mandated system of any similarity. Also of note, the results of over 28,000 pregnancies which lasted longer than 43 weeks were excluded due to coding problems—a reasonable enough number that could have had an influence on the overall results, had they been included.
The researchers aimed to relate maternal age ranges (five year increments between 20 and 45) to various weeks of gestation (16-22, 23-29, 30-36, ≥45) and, thus, determine “the association of fetal death with maternal age by length of gestation.” In general, the results suggested that increased maternal age + increased weeks of gestation (>40yrs. old, >36 completed weeks of gestation) resulted in a significantly increased risk of stillbirth. The resultant jump in risk of fetal death at term from >40-year-olds to >45-year-olds is almost two-fold. Despite this, and other studies showing similar results, “Advanced Maternal Age” tends to include women aged 35 and above, rather than starting the definition at age 40. (Recommendations on how to manage “higher risk” pregnancies in women >35 y.o. are also linked to well-documented increased risk of chromosomal birth defects as maternal age advances beyond 35.)
Admittedly, the data this large study provides is worthy of great attention. However, before implementing practice guidelines based on these findings, such as “All women aged 40 or older should be disallowed to carry a pregnancy beyond 37 completed gestational weeks,” I believe several other factors need to be analyzed.
A Closer Look
We know that multiple underlying conditions, beyond those controlled for in the Haavaldsen study, are associated with increased rates of preterm birth and/or stillbirth, some of which include: pre-pregnancy diabetes (and gestational diabetes), pre-existent thyroid disease, obesity and smoking. Race, unfortunately, also plays a statistical role. And yet, none of these potentially underlying factors are addressed in the study. In fact, we have no idea how many of the study participants may have had one or more of these potentially confounding variables. (This very issue is, interestingly enough, brought to light in the ACOG Guidelines for Stillbirth Management.)
Likewise, other factors that have not garnered much attention in the literature but, in my estimation, certainly influence a woman’s general state of health and well-being (and thus potentially, the health of her pregnancy) are factors such as: diet, exercise routine and overall stress level. Designing a future study which could control for these additional variables would undoubtedly alter the data tremendously, prompting questions such as, “to what exponential degree does any one underlying condition increase the relative risk of fetal death?” Likewise, assuming that the Norwegian study participants may have had some of the additional variables I suggest here, in some cases relative risk of stillbirth could even be decreased in any given maternal age group, when that/those underlying conditions were controlled for. Following such a study, would it be feasible then, to develop an Adjusted Relative Risk (ARR) Coding System for pregnant women which would go on to advise certain levels of perinatal and intrapartum care, based on the identified code?
Let’s take a look at a few hypothetical pregnant women to prove my point, but through a more intuitively analytical lens:
|Sample Pt.||Age||Gest. Wks.||BMI|| Smoker
What type of relative risk of stillbirth would these women likely have?
I think it is fair to say that, assuming all other potentially confounding variables are controlled for, Woman A has the lowest ARR and Woman E has the highest. This is where statistics from studies like that of Haavaldsen, et al come in to play. But what of the other women?
Woman B falls into the age range and gestational weeks that earns her an ARR of 1.00 (very low) in the Haavaldsen study, but how would that risk look differently when we consider her BMI (indicative of obesity) and her positive gestational diabetes status? She’d have a higher risk of stillbirth, right?
Woman C presents a similar quandary: If, again, all other potentially confounding variables had been controlled for, her age and gestational weeks would place her in a relatively higher state of risk compared to Woman B (ARR would be 1.60 in the Haavaldsen study) but would this be an accurate estimation of her stillbirth risk? What about her smoking and thyroid disease status? Wouldn’t these earn her an even higher ARR, suggesting the need for an even greater degree of attentiveness by her health care provider?
Woman D presents a loftier challenge, still. At 41-years-old and 41 weeks, 5 days pregnant, the Haavaldsen study would lead us to automatically believe she is at a relatively high risk for stillbirth as each day of her pregnancy progresses. But what of her other health factors? Does the absence of obesity, smoking and thyroid disease discount some of the risk her age and length of pregnancy create?
As a childbirth educator, I witnessed the angst provoked by our (currently limited) categorical risk assessment. I recall one woman in particular—a 43-year-old first time pregnant woman who was rapidly approaching her due date. Her healthcare provider—offering guidance from the most current research, I’m sure—was working hard to get her to agree to a labor induction. Having prepared extensively for a normal birth, she was extremely frustrated by her doctor’s attempts to treat her as an at-risk mother of an at-risk baby.
“I feel fine,” she kept telling me. “I feel healthier than I have in years. Why do they want to induce me when nothing is wrong?”
This woman was an avid hiker and skier, she had a low-stress job, she ate healthfully, didn’t smoke, had no underlying medical conditions and appeared to have been of normal weight (prior to and during pregnancy).
I found myself, her Lamaze class instructor, stuck between a rock and a hard place. I understood (I think) the impetus of her doctor—recognizing statistical trends, she wanted to save her patient from the potential heartache of a stillborn baby. And, yet, this mother’s intuition kept telling her everything was fine with the baby, as well as with her own health status.
What really was the most appropriate approach to this woman’s care?
Other Factors to Consider
Other elements of the Haavaldsen study I feel compelled to mention include the suggested role labor inductions played in the cohort:
Term predicted by ultrasonographic examination may, for many obstetricians, be more trustworthy than term predicted by last menstrual period. Hence, induction of labor may have been more common in pregnancies estimated to be postterm by ultrasonographic examination, than by [LMP].
Despite the mention of induced labors being included in the cohort, there are no numbers on how many of the 2 million + pregnancies ended in induction—leaving a potentially HUGE confounding variable unchecked.
Lastly, I appreciated the authors’ mention of the role cervical cone excision may have played in the early pregnancy stillbirth risk as well as preterm labor risk. As the researchers described,
The relative risk of preterm delivery associated with cervical cone excision, increases with decreasing duration in pregnancy. In preterm deliveries, fetal deaths are likely to happen during labor and are caused by immaturity.
The Haavaldsen study included a review of 37 similar works of research which, according to study authors, “demonstrated a significant increased risk of stillbirth in women of advanced maternal age.” Because of that, (but despite the fact I, personally, hate the term “Advanced Maternal Age”) I don’t believe we can discount any and all motions to induce labor in women who fall into the highest of high risk categories. But I also believe there is much more work to be done to further validate this data in terms of the multiple other risk factors associated with stillbirth. Until we can complete this picture, in my mind, the jury is still out.