In September of 2019, The Lancet published a cohort and meta-study analysis titled, Dietary carbohydrate intake and mortality: a prospective cohort study and meta-analysis. The researchers suggested that a low carbohydrate diet was associated with shorter lifespan than those whose consumed a plant-based diet.  Longevity has not been studied in relation to a low carb ( 100 g/day), very low carb ( 20 gm/day), or no carb study, but it would stand to reason that if eating this way affected longevity, our race would have died out a millennia ago. To really understand what the science is saying, it’s important to review the actual study and in this case, there were several red flags.
There are 2 parts to this study, the cohort group and the meta-analysis.
Definition of Low Carb
While the study stated the participants ate low carb, their diet, which consisted of around 40% carbohydrates or 244 g, is not low carb. It is, however, lower carb than the standard American diet which is 45-65% of calories. Most low carb eaters eat less than 20 g of carbs or 5% of total calories. At 40% of calories the body is still burning carbohydrates for fuel and storing the rest. A diet 40% carbohydrates will store the carbs and fat eaten are stored as fat.
The other question that comes up when thinking about this study is what kinds of food comprised the 244 g of carbs. Highly processed foods with sugar, starches, and trans fats? Pastas? Bread? Fried Foods? Soda? Seed Oils? None of which are heart healthy.
Control Group 
Another cause for concern is that the study portion of this trial had no control group to adjust for unhealthy habits, exercise, or even stress. Control groups allow researchers to change conditions in the treatment group or in this case, the cohort group while keeping the control group constant. Scientists are then able to compare results.
Preliminary Participant Data Collection
In the cohort portion of the study, the researchers adjusted for income level, exercise, cigarette smoking and diabetes. The authors of this paper excluded the definition of exercise. By reading the study, one cannot be certain of the type, how much, how long, or the percentage of participants reported physical activity. It is also not clear if someone was a non-smoker at the start of the study if they remained smoke free throughout the study period, or did they start smoking during the study. The scientists identified people with diabetes if 1) the participant had heard their doctor say they had the condition, 2) the participant took diabetes medication, or 3) the person had a fasting blood sugar 126 mg/dl or a non-fasting blood sugar 200 mg/dl. It is important to note that a physician would not make a diagnosis of diabetes based on a single sample. Furthermore, no mention is made of pre-existing conditions such as stress, prediabetes, hypertension, heart disease, stroke, or other chronic conditions which are indicators of a sick metabolism.
The Lancet study also did not mention if the studies included in the meta-analysis were adjusted for chronic conditions, inactivity, or cigarette smoking.
Observational studies “suggest” association, but do not prove causation. In the cohort analysis of this trial, data were collected by asking people to complete a food frequency questionnaire every 6 years. This means that participants needed to recall how much of what food they ate on average over a year. Without a food log, it would be difficult to accurately state what someone ate last month let alone 6 years ago.
- Visit 1: 1987-1989. Participants recruited
- Visit 2: 1990-1992
- Visit 3: 1993-1995
- Visit 4: 1996-1998
- Visit 5: 2011-2013
- Visit 6: 2016-2017
To assist their recall, cohort participants picked foods from a list provided by researchers. If something they ate in the last 6 years was not on the list, it was not included in the study. I question if the lists were even updated from year to year with new foods. While there may be a small percentage of people who are good at recalling what they ate during the last 3 or 6 years, most of us would struggle answering the questions.
Correlation versus Causation 
Ketogenic.com also reviewed this study and published a list of their own concerns, describing correlation and causation this way:
“This study is purely correlational, and correlation does not equal causation. Think of it like this: If a new study was published showing individuals who wear purple socks were more likely to get into a car crash than individuals wearing red socks, would you assume that purple socks cause car accidents? You probably wouldn’t and the same principle applies to this study.”
The Lancet study contains enough red flags for the reader to dismiss the conclusions. When in doubt, read the study yourself and ask, “are the people in the study like me.”