Skip navigation

Testing Wrap-up: Rollerskiing and insulin dependency

Kris Freeman

Tue, May  20, 2008 - By Zach Caldwell

Today Kris finished a tough three-day testing camp in which he went to max on the roller ski treadmill on all three days. Day one was the standard sub-max (lactate profile) protocol followed by a max VO2 stage. As I’ve already reported, that test went well - Kris had his second best test ever on this protocol - just behind his effort last October when he was in very good shape. Yesterday and today were a bit different.

VO2 Max testing and lactate profile are interesting benchmarks, but they seldom help steer training or racing strategies in a really intelligent way. Even the establishment of training “zones” is largely wasted time with elite level athletes (this is sure to raise some hackles - so it’s worth pointing out that this is just my point of view). Last year Kris’s July test showed a downturn in his peak capacity. In hindsight it looked an awful lot like he was in trouble - walking a fine-line between a good high-volume load and a dangerous overtraining load. I’m sorry to say that we were able to rationalize that test result - you can look back in the archives at what I wrote if you’d like. We weren’t stupid - we had our eyes open and we made the decisions we made with full consideration. But the test didn’t stop us from accompanying Kris further down the road toward an unstable and unsatisfactory season.

I would still favor getting regular treadmill data on athletes. There is a great deal of information that can be mined out of the raw numbers - particular with respect to substrate utilization and lactate mobilization. This information can certainly help in the planning process, but to know how an athlete will respond to training you’ve still got to work with the athlete. And all the max tests and lactate profiles are a much more useful way to consume a coach’s attention than to benefit an athlete.

For Kris the tests we ran yesterday and today were a bit different from the standard deal. We had specific goals for starting to build a better understanding of his insulin dependency. Several years ago we noticed a correlation between his best race performances and relatively high basal doses of lactate. When the training load is high, Kris’s insulin sensitivity is enhanced and he doesn’t need to supplement very much insulin in order to maintain his blood sugar levels in the appropriate range. As the training load comes down his insulin sensitivity decreases and he has to take more insulin. His basal dose is the amount of long-acting insulin that he keeps in his system to maintain a systemic equilibrium. He’ll have to inject short-acting insulin specific to the type and quantity of calories that he consumes, but he’ll always have a baseline level that reflects his baseline sensitivity.

When we recognized the relationship between good performances and high basal doses we naturally assumed that Kris’s basal dose was acting as a barometer for his state of recovery. When he’s more recovered his insulin sensitivity goes down so he’s taking more basal insulin, and at the same time his performance gets better because he’s rested. Seems logical. We also figured that there was a certain danger to using too much insulin as it is likely to result in low blood sugar - a state that is definitely not conducive to racing fast.

However, over time we started to see evidence that Kris’s greatest blood-sugar-related difficulty in race situations came with high blood sugar. Once in a while he would end up low, and out of energy, but generally only when he missed feeds. High blood sugar - the result of under-dosing insulin - on the other hand, was more confounding. It seemed to be correlated with just plain sub-par performances, and when we got lactate data the numbers were always alarmingly high. This started us thinking along a couple of different lines from our previous model. Most obviously, there appeared to be a correlation between blood sugar and blood lactate - that’s intuitive, and normal in non-diabetics as well. More importantly, we started thinking of insulin as a mechanism for moving sugar - making it available as fuel. For somebody involved in endocrinology this might be really basic stuff, but we’re not quite so worldly. Heck, you can most of this in the first paragraph of the wikipedia entry on insulin, but it wasn’t an intuitive leap for us to make.

Last Summer Kris started experimenting with higher insulin doses on race days. He had always increased his basal dose a bit because the adrenaline involved with race day would cause an increase in blood sugar. However, increasing his dose for a race was a tricky thing for a major logistical reason. He had two types of insulin to work with - long acting insulin which he has to inject well in advance and which stays in his body for up to 24 hours, and short acting insulin that does its work and dissipates very quickly. If Kris wanted to jack his long-acting dose up quite high for a race effort, it means that he’d have to consume a tone of carbohydrates for the entire time that the long acting insulin was in his system. Sleeping could be dangerous, as his blood sugar levels would drop. There were obvious limitations to his management strategy.

In April Kris started working with the Omnipod. It’s a very compact insulin pump that sticks to the back of his arm with adhesive. It is programmed wirelessly with a PDA-like device which doubles as a glucometer. The Omnipod uses only short acting insulin, but it feeds him a tiny dose every five minutes. He can program it in advance to change the dose every 30 minutes. Because he has no long-acting basal dose in his system, he can control his insulin dosage quite well over the course of the day. This opens the door to very different race-day insulin dosing strategies. In fact, it completely removes the practical limits to insulin use that he was operating under previously.

The testing that Kris did yesterday and today was designed to start to explore the situation more completely in a controlled environment. First, we wanted to know whether we could document an improvement in his performance with an increase in his dose. We also wanted to begin to probe the limits of this model. And we also wanted a better peak inside his system at what is actually happening with his blood sugar and his lactate concentration during a race effort. While we were all in Whistler Kris and Pete and I sat down with Randy from USST sports science to design some tests to open start to answer some of these questions. The test we came up with was a simulated race effort in which Kris skis three five minute stages on the treadmill at a workload calculated based on his submax protocol to be just beyond lactate threshold. after each stage he takes a short break and the lab techs take a finger-prick blood sample for sugar and lactate. The fourth stage starts off just like the first three, but it’s progressive, with the treadmill getting steeper at regular intervals. The idea is to take Kris to failure. Since we want to measure workload capacity we felt that a progressive protocol taking him to failure would give us the best picture of what he’s actually able to do. The gas exchange data is an extra bonus, not the basis for evaluation of his effort. If a given dosing strategy allows him to go significantly further but at a lower VO2 we’re not going to call it a worse test!

The results of three days of testing were very interesting. I haven’t got the raw data yet, but we saw some unexpected things. One of the factors we were interested in testing was whether the correlation between sugar and lactate held up at low to moderate workloads. In the warmup for his day-one submax test Kris started skiing and at his first check his blood sugar was quite high (he figures he was underdosed because of the adjustment to altitude which always demands additional insulin from him). His lactate at that point was 2.3 mMol/L, uncharacteristically high for such an easy load. He programmed a higher dose into the Omnipod, and within minutes, while he maintained exactly the same treadmill workload, his sugar came down to a normal range and his lactate came down to 0.8mMol/L. This appears to indicate that the sugar/lactate relationship holds up at lower intensities quite strongly, and the interesting thing is that the high sugar level, while high, was not quite in the range where Kris identifies heavy performance issues. This is good reason to keep an eye on both sugar and lactate levels during even easy training, as carrying lactate concentrations in the 2-3mMol/L range is an unnecessary stress on the system.

Another interesting outcome of the testing is that, under all dosing scenarios - even the high one, his blood sugar increased as he cranked the effort up toward max. It appears that his body dumps sugar into the system as the effort goes up. The interesting thing is that a similar thing happens with Andy Newell who submitted himself to blood sugar testing as a control. However, Andy’s increase was much smaller, and all within a very healthy normal range. It appears that the sugar-dump might be totally normal, but that Andy can regulate it on the fly - effectivelyproducing insulin and utilizing the sugar immediately. The implication is that the required “normalizing” dose might increase during a sustained progressive effort.

Kris’s highest dosing scenario produced his best result. Not by an alarming margin - it didn’t turn him into a superman. But he stayed on the treadmill longer, reached a higher peak VO2 and produced less lactate. His blood sugar numbers in that final high-dose test showed the same upward trend (with a small dip on one stage), and very closely mirrored Newells sugars. However, when he first got on the treadmill Kris’s blood sugar was at 39 - a level that would cause many people to pass out. He had started the high dose a while in advance of the test, and it had certainly gone to work by the time he started. He consumed 20 oz of gatorade and started the effort.

There is a great deal left to learn about how to optimize Kris’s insulin dosing strategy. I won’t even try to enumerate the issues that might emerge as he moves forward. Meanwhile, Randy is already talking about how to design the July testing session to best effect, and Kris has an improved understanding and some powerful new tools at his disposal. This has been an extremely fruitful testing camp.