Menstrual cycle length… What’s normal?

5 July 2018.

Image credit: Pastorius on Wikimedia- 

Many people are concerned because they think that their cycle is too short, too long, or just not ‘regular’ enough...

Nabta Playa Calendar Circle, reconstructed at Aswan Nubia museum, Egypt

Sometimes, there is a health issue behind an unusual cycle length, but more often than not, this is just unnecessary anxiety… The result of yet another menstrual myth- the ‘myth of the 28 day cycle’.

It turns out that we have been misinformed about what counts as ‘normal’ cycle length for THOUSANDS of years!

We can excuse Ancient societies (who generally had a fairly loose relationship with time and calendars) but how does this myth persist, in a world full of calendars, smartphone applications, and the internet?

1. We are lazy when it comes to maths…

Just like the ‘menstrual synchronicity myths‘, the concept of the normal ’28 day cycle’ comes from a mathematical misunderstanding. 28 days is actually just the ‘average’ cycle length, of the vast majority of menstruating people from approximately 20 to 45 years old [1]- who are not using hormonal contraceptive medications or devices, and do not have a health condition that can affect cycle length e.g. PCOS (Polycystic Ovarian Syndrome), anorexia, extreme weight loss or weight gain, thyroid conditions, or early menopause.

Actually, the average menstrual cycle length changes with age; at age 20, the average cycle length is typically around 29.8 days, by age 40 the average cycle length has decreased to around 27 days [2]. Before age 20, the hormonal changes associated with puberty result in far greater irregularity (resulting in an average cycle length of around 35 days) and similarly, as people approach the menopause, the average cycle length abruptly extends to over 44 days. [2]

The thing is, mathematical averages are not always the most useful way to describe something, especially when that thing (in this case ‘normal menstrual cycle length’) is actually a range between 21 and 40 days. ANY cycle length between 21 and 40 days long is considered NORMAL [3].

What’s more, an ‘average’ cycle length calculation (which is always a single number) hides the fact that people who menstruate typically have variable cycle lengths [1, 2, 5]! That’s right, variation is totally NORMAL (usually between + or – 4 days from the average cycle length for that individual) [6].

In fact, it is so rare to have exactly the same cycle length for more than three consecutive cycles that menstrual health researchers have a special name for it- ‘isochronal’ cycles [7]. Such regular menstrual cycle length can even be a sign that a research participant is perhaps not keeping accurate records [8]!

Graph showing typical patterns of cycle length variation [5]

2. Human nature gets in the way of accuracy

Besides this reluctance to properly ‘do the math’, we humans also tend to be a bit lazy about ‘evidence-based thinking’ in general [4, 9]. For example, we much prefer to hear about, remember, and share ‘simple’ or ‘tidy’ ideas to those that require a bit more nuance (subtle and careful description) [9].

This tendency to simplify information into easy to remember ‘facts’, was probably very helpful during human evolution, when we needed to make quick decisions in potentially life-threatening situations. However, it is less useful when it comes to communicating accurate information about menstrual cycle length;

Another major contributing factor (also related to human laziness!), is that very few people actually bother to track their cycles with much precision, and even those that do, rarely take the time to calculate and compare cycle lengths over time [10]. A lot of people rely on their memory to jot down the first day of their previous period- so mistakes are quite likely, especially if more than a few days have passed…

It is usually only when people are trying to conceive a baby that they (and/ or their partner) get serious about tracking their menstrual cycle. This is why so many people who think that they have irregular cycles, start tracking, and discover that they have a 21-40 day NORMAL cycle with natural, and NORMAL variation! Similarly, some people are convinced that their periods are “regular as clockwork”, until they start to track their cycles [10]…  Sometimes they then panic and think that something must be wrong!

Medically speaking, an ‘irregular cycle’ is defined as being one that suddenly changes (without hormonal medication being involved in any way), or is typically longer than 40 days, or involves mid-cycle bleeding– it does not refer to natural variation in cycle length within the normal 21-40 day range.

3. Gender inequality (especially the menstrual taboo) creates silence

Since this evidence-base has been around since at least the 1930’s [1], it seems strange that the ’28 day cycle’ myth is still so powerful today… It is therefore likely that social and political factors have prevented this information from becoming more widely available and known.

As previously discussed, the menstrual taboo (the ancient idea that the menstrual cycle is somehow shameful rather than a sign of reproductive health) results in the silencing of open discussion of this topic, even within the healthcare setting.

For example, fertility researchers have had to come up with complex mathematical models to deal with cycle length variation (not only due to individual’s natural variation, but because the overall variation in the population is not ‘normally distributed’- due to cycle length changes, especially at puberty and during the menopause [11]). However, this ‘common knowledge’ of cycle variation has not filtered down, or been adequately communicated to, the public, or even non-specialist clinicians.

Related to this issue is the current (and historical) lack of adequate menstruation education. In fact, I am yet to find a biology textbook about the reproductive system that includes information about natural cycle-length variation. This means that in the UK, in 2018, people who menstruate are still almost NEVER TOLD that although 28 days is the average cycle length, this does not make it a useful description of what counts as ‘normal’.

Finally, this myth also reinforces the idea that ‘all women’ are somehow biologically ‘the same’, as outlined in an earlier blog. This is a common way to ‘naturalise’ unequal power relations. By generalising in this way, it is possible to attribute the particular characteristics of any one individual to the entire group, usually as a means to discredit or undermine their relative position in society. This is a key part of racism, sexism, xenophobia, religious hatred, and homophobia. By maintaining the myth of the ‘regular 28 day cycle’, regardless of overwhelming evidence that shows  variation is the norm, we are perpetuating this harmful ideology.

References and notes:

1. The largest and most rigorous ‘classic’ study on menstrual cycle length was carried out way back in the 1930’s- Gunn, D.L., Jenkin, P.M. and Gunn, A.L. (1937) Menstrual periodicity: statistical observations on a large sample of normal cases. J. Obstet. Gynaecol. Br. Emp., 44, pp 839–879

In more recent research around menstrual cycle length the same patterns of individual variation, and decreasing average cycle length with age (until peri-menopause), are consistently found in studies of populations from around the world e.g.

  • Jeyaseelan, L., Antonisamy, B. and Rao, P. S. (1992) ‘Pattern of menstrual cycle length in south Indian women: a prospective study.’, Social biology, 39(3–4), pp. 306–9. Available at: (Accessed: 5 July 2018)
  • Harlow SD, Lin X, Ho MJ. (2000) Analysis of menstrual diary data across the reproductive life span applicability of the bipartite model approach and the importance of within-women variance. J Clin Epidemiol. 53(7):722-33
  • Creinin, M. D., Keverline, S. and Meyn, L. A. (2004) ‘How regular is regular? An analysis of menstrual cycle regularity’, Contraception, 70(4), pp. 289–292. doi: 10.1016/j.contraception.2004.04.012
  • Ferrell, R. J. et al. (2005) ‘Monitoring reproductive ageing in a 5-year prospective study: aggregate and individual changes in steroid hormones and menstrual cycle lengths with age’, Menopause, 12(5), pp. 567–577. doi: 10.1097/01.gme.0000172265.40196.86
  • Guo, Y. et al. (2005) ‘Modelling menstrual cycle length using a mixture distribution’, Biostatistics, 7(1), pp. 100–114. doi: 10.1093/biostatistics/kxi043
  • Cole, L. A., Ladner, D. G. and Byrn, F. W. (2009) ‘The normal variabilities of the menstrual cycle’, Fertility and Sterility. Elsevier, 91(2), pp. 522–527. doi: 10.1016/j.fertnstert.2007.11.073.

2.  These particular averages are taken from Vollman, R. F. (1977) The Menstrual Cycle New York; Knopf pp 51-52 but again, they are consistent with more recent high quality studies [see note 1- above]

3. As defined by the NHS (2016) Periods and fertility in the menstrual cycle. [ONLINE] Available at: [Accessed 5 July 2018]

4. I cannot recommend this book highly enough. Sadly, Hans Rosling died last year, but his work continues to inspire millions of us who want to change the world! Rosling, H., Rosling, O. and Rönnlund, A. R. (2018) Factfulness : ten reasons we’re wrong about the world – and why things are better than you think Sceptre: London pp 39-40

5. Treloar, A. E., Boynton, R. E., Behn, B. G., & Brown, B. W. (1967) “Variation of the human menstrual cycle through reproductive life” International Journal of Fertility 12: 77–126

6. Cole, L. A., Ladner, D. G. and Byrn, F. W. (2009) ‘The normal variabilities of the menstrual cycle’, Fertility and Sterility. Elsevier, 91(2), pp. 522–527. doi: 10.1016/j.fertnstert.2007.11.073

7. The earliest reference (I’ve found) to ‘isochronic cycles’ is in Vollman, R. F. (1977) The Menstrual Cycle New York; Knopf pp 50

8. For example, this issue is mentioned by Treloar, A. et al. (1967) [see note 5 above], Creinin M.D. et al (2004) [see note 1 above] and Small, C et al. (2007) [see note X below]

9. Kahneman, D. (2011). Thinking, fast and slow. New York: Farrar, Straus and Giroux.

10.  Small, C. M., Manatunga, A. K. and Marcus, M. (2007) ‘Validity of Self-Reported Menstrual Cycle Length’, Annals of Epidemiology, 17(3), pp. 163–170. doi: 10.1016/j.annepidem.2006.05.005.

11. For example, Guo, Y. et al. (2005) ‘Modelling menstrual cycle length using a mixture distribution’, Biostatistics, 7(1), pp. 100–114. doi: 10.1093/biostatistics/kxi043

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.