A lot has been made of “spinning” and keeping a high cadence in the last 10 or so years. Of course, many of us have heard the story about how Lance Armstrong optimized his riding style by changing his average cadence and in the process made himself into a grand tour contender.
Here is the basis of it: higher cadence has a lower muscular cost, and a higher cardiovascular cost. Lower cadence is a higher muscular cost, and lower respiratory cost.
When the muscular system is overly taxed, like in a hard one day bike race where you rode with a high wattage, but a low cadence, it can take a few days for the muscular system to break down and repair the damaged tissue. If you're just doing a one day race and won't race again for a few days to a few weeks, then this may be just fine, and as we will find out later it may be the best game plan to do well in that race.
If you're riding many days in a row, as in a grand tour (3 weeks) or a shorter stage race (say, 5 days to a week long), you won't want to carry this muscle damage from day to day, since recovery will be far from complete, and in fact, you could experience progressive breakdown, and a precipitous decline in performance.
The higher strain on the respiratory system has much shorter term down side. The respiratory system, our lungs and heart, can recover after a hard ride within a few hours. Think about that: imagine you just did a hard ride (at any cadence) that was an hour longer than any ride you've done so far this year. Your lungs may be a little “phlegm-y” for a few hours from breathing hard, but it's usually gone by the evening or in the morning. Your legs, however, can be sore for many days after.
The more consistently you ride, and the more consistently you race, the more useful a high cadence is. If you ride only a couple days a week or, more importantly, compete only infrequently then there is not much benefit to pedaling with a cadence over 90-95 rpm.
As with most things, “low” cadence is also a matter of degree. I would consider 90-95 rpm “high”. When the magazines and websites began touting the benefits of high cadence, many cyclists, as is common for their type-A, if-some-is-good-then-more-must-be-better attitude, immediately decided if 90 was good, then 110 must be even better. I had, and still have, dozens of bike fits every year where the client is spinning madly, in an obviously uncontrolled way to reach their high target. Things brings me to two points:
Extremely high cadence: 1. requires more coordination and pedaling skill and, 2. makes small mechanical deviations more likely to create injury
Doesn't extremely low cadence can also create injury? It may, but I can say from two points that cadence on the lower end requires much less coordination and skill and it does not make mechanical deviations (like a drifting knee) worse overall – this has been born out over years of performing bike fits. Remember those clients spinning madly on their bikes? Well, their 3D motion capture numbers were often all over the place: Knee Lateral measurements over 50-60mm each side, Knee Angles anywhere from 6 - 15 degrees, Hip Vertical Travel measurements excessive, etc. etc. I would then make them shift into a bigger (harder) gear, and take another view with the infrared -- voila! Usually about a 30% reduction in the aberrant motions. Then we were able to get down to the business of finishing their fit.
Studies have varied, but the median for overall efficiency is somewhere in the low 80s, which, in my experience, fits with the clients who have the lowest incidence of injury. (Again, this is an average or a generalization over hundreds of bike fits, but indicative nonetheless.)
So should you grind up all your climbs at 45 rpms? Of course not, but there is no good reason to spin away at 110 rpms on the flats either.