Fasting and Sleep: A Retrospective Analysis
Fasting and Sleep: A Retrospective Analysis
As most of our readers know by now, sleep is one of my favorite topics. I have written about it extensively in the past, including my four-part series on sleep, and my review of sleep tracking devices. Over and over, I see many articles written on how intermittent fasting can help improve circadian rhythms and sleep quality. However, despite my searches for clinical research on time-restricted eating, especially surrounding Ramadan, I found nothing on humans regarding extended fasting and sleep.
This made me wonder if I had personally experienced any of these effects, and what the reports were based on. Considering I have been wearing the Oura Ring daily since implementing fasting into my daily routine, I decided to let the numbers speak for themselves.
First, I evaluated my tracked sleep for 30 days when I started fasting, and then again after a year of fasting. I then decided to add a third set of data: corrected 2019 results removing every night during which I knew I had been heavily intoxicated (there were no nights in the 2020 period in which I was heavily intoxicated). Here are the results:
2019–2020 Sleep Results
|Data||Sleep time in minutes||Sleep Score||Rem time in minutes||REM %||Deep time in minutes||Deep %||HRV in ms||RHR in beats/min||RR in breaths/min||Body Temp in celsius|
HRV = Heart Rate Variability
RHR = Resting Heart Rate
RR = Respiratory Rate
Deep = periods of deep sleep
Blue is “most favorable,” and red is “least favorable”. (i.e., it is more favorable to have a greater amount of total sleep, sleep score, REM time and %, deep time and %, and a higher HRV). It is more favorable to have a lowered heart rate and a lowered body temperature. All respiratory rates were within “normal,” so evaluating the most “optimal” was not possible.
Statistics are funny. It’s not surprising that there was no statistical significance, but definite practical and clinical significance, due to having data from only one participant and not many data points; however, HRV and RR were approaching significance at p=0.055 and p=0.058, respectively.
The difference in respiratory rate between groups was largely insignificant from a practical standpoint, although the slight increase in HRV from 2019 to 2020 may be significant. That said, the corrected 2019 results saw me with the most sleep time, the highest sleep score, the most REM time, as well as highest REM %, Deep time, Deep %, and also with the lowest resting heart rate and the lowest body temp drop, indicating better recovery. It seemed that overall, my sleep quality had slightly decreased over the year of fasting. This was despite losing approx. 45–50 pounds during this time as well, which should have come with its own sleep quality benefits.
Next, I decided to evaluate my sleep during various time points of the fast against non-fasting controls. These results were far more significant. My editor, who conducted the statistical analysis, made a couple of observations:
“The one-way ANOVA analysis revealed significance between the four groups (No fast, 24-hour fast, 48-hour fast, and 72-hour fast) for sleep time, REM time, and RR.
Even though significance was not achieved for the other results tested with ANOVA, the p values were still on the small side, approaching significance.”
One caveat of this analysis is that it was not feasible to accumulate large data points for each situation, as there are only so many 48-hour+ fasted before bed nights, and especially 72-hour+ fasted before bed nights I can conduct in a healthy manner. There were 20 data points for no fast controls, 16 for 24-hour+ fasts, and 3 of each for 48-hour and 72-hour fasted before bed nights. Here are the results:
Sleep Results After Fasting
|Data set||Sleep time in minutes||Sleep Score||REM time in minutes||REM %||Deep time in minutes||Deep %||HRV in ms||RHR Beats/min||RR breaths/min||Body temp in Celsius|
|24hrs fasted before sleep||389.75||74.31||74.375||18.94||120.44||31.19||79||45.63||13.89||-.02|
|48hrs fasted before sleep||348||67||69.33||20.33||111||32||78.33||45.67||13.87||-0.2|
|72hrs fasted before sleep||273||66.33||38.33||13.67||106.67||40||58.33||49.67||15.1||-0.5|
Next, my editor conducted analysis within the groups:
“To further investigate which of the results significantly differed from each other, I carried out a series of Tukey-Kramer post-hoc tests. The significance is marked with an asterisk in the graphs.”Graphs of Sleep Results After Fasting
Overall, it seems that extended fasts of 72 hours have significantly impaired sleep, and the “no fast” controls have the best sleep. That said, I often feel the most rested after fasting for 24 hours prior to sleep. It is something I have anecdotally noticed, particularly on days when I have properly transitioned into a fast by eating healthy the day before. While not significant, there seems to be a J curve at play, where the 24-hour and 48-hour fasts reduce heart rate.
Significantly, respiratory rate drops during this period as well, and temperature slightly dips. For 24-hour fasts, the drops in other parameters, such as sleep score, REM, and HRV, are marginal and insignificant. Moreover, the total sleep time is reduced by 26 minutes, and deep time goes down by about 15 minutes; but at the same time, there is an increase in the proportion of sleep time spent in REM (REM%).
It is possible that a properly executed 24-hour fast can have a negligible, or even positive, acute effect on sleep. As it is with other forms of hormesis, the “dose makes the poison or benefit,” and this would likely vary widely from individual to individual.
Effects of Extended Fast vs. Extreme Intoxication on Sleep
Finally, in reviewing the significant sleep impairment observed at 72+ hours of fasting, I wanted to evaluate the differences between sleep quality seen after the fast, or after extreme intoxication (nights when there was a loss of memory, for instance). Currently, I am collecting more data on alcohol’s role on my sleep at several different controlled intake levels. Here are the results:
Sleep Results of Extended Fasting and Extreme Intoxication
|Data Set||Sleep Time||Sleep Score||REM time||REM %||Deep Time||Deep %||HRV||RHR||RR||Body Temp|
|72+ hour fast||273||66.33||38.33||13.67||106.67||40||58.33||49.67||15.1||-0.5|
From my experience, I am not sure which leaves me feeling worse: having fasted for 72 hours prior to a night’s sleep or being heavily intoxicated prior to sleep. Both leave me feeling quite lethargic and exhausted. During extreme intoxication, I get more total sleep and more REM sleep, but my deep sleep, HRV, resting heart rate, and body temp are all more impaired. It was interesting to me to see the differences, and how each impaired my sleep quality in a different way.
I’ve read and heard many times that alcohol reduces or even completely blunts REM sleep. I have always questioned this, as many times over the many years of my adult life I have awakened from severe intoxication, sometimes still intoxicated, and have not remembered getting to bed, or much of the night, but have had vivid memories of my dreams.
My editor commented on the statistics:
“I had also carried out a series of T-tests as you had requested to test between 72-hour fast and severe intoxication groups. Some of the results were significant:
-Deep time was significantly lower in severe intoxication (this makes sense, as alcohol interferes with sleep).
-HRV was significantly lower in severe intoxication (this also makes sense, as alcohol is a muscle relaxant and nervous system depressant).
-Body temperature was significantly higher in severe intoxication (also makes sense due to alcohol’s vasodilatory effects on the vascular system).
-Deep % results approached significance, but significance was not achieved (p=0.06).”
Here are the results:
Graphs of Extended Fasting and Extreme Intoxication Results
Fasting and Sleep: The Takeaway
I’m glad I have evaluated this data on fasting and sleep. I believe this further emphasizes the point that individuals, if wishing to maximize their own health experience, should be recording and evaluating the data they have available, especially in regards to using hormetic stressors (i.e., alcohol, exercise, fasting, cold or heat exposure, and perhaps even molecular hydrogen).
Keep on the lookout for the launch of my not-for-profit platform for the general public and researchers alike to submit this type of data and take advantage of the collective knowledge. Tutorials, templates, etc., will all be available.
This is a project I have launched, but many others are assisting with it.