DIY sweat composition testing

@PrecisionHydration any feedback you have on this approach would be appreciated. Thanks!

I’ve been interested in getting a sweat composition test, but the cost has put me off, so I’ve looked at how I could do something reasonably accurately myself.
What I have come up with is the following plan, which I’ve tested a few aspects of allready.

For measuring the sweat, I will be using a HM Digital AP-1 total dissolved solids tester (https://smile.amazon.com/HM-Digital-Dissolved-Resolution-Accuracy/dp/B007KDYOMU?sa-no-redirect=1) This the main cost of the experiment, at $24. This measures up to 5000 ppm NaCl. Not that it doesn’t tell you what is in the water, just how much - it does this by measuring how conductive the water is. This unit is calibrated to an NaCl standard, so 1 gram of NaCl in 1 Kg of water will read 1000 ppm. You need a fair bit of liquid (a few ml at least) to measure, so while this can measure sweat directly you have to collect a lot of it.

From the perspiration wikipedia page, the linked paper ([PDF] Sweat mineral-element responses during 7 h of exercise-heat stress. | Semantic Scholar) shows that sweat is about 80% sodium, 20% potassium, with the rest all trace amounts.

I will be collecting the sweat from my forearms, as this is easy to access, and there is data on how to convert concentration here to overall body concentration. Since I need to collect a lot of sweat, I will be using paper towels over both forearms to collect enough sweat. Initial testing showed that paper towels added some dissolved solids to the water, so for the real test I have washed/dried paper towels in reverse osmosis water to minimize this.

The Gatorade sports science group has a nice page about sweat testing ( www.gssiweb.org) and this includes an equation for calculating whole body sodium loss from measuring 1 specific site - the forearms. After converting that into units that I am measuring, I get:
Predicted Whole Body sweat [NaCl] (g/L) = 0.57 (forearm sweat NaCl) + .618
What is interesting here is that the whole body sweat concentration is a fair bit less than the concentration measured on the forearms.

My plan is to do the following:

  • wrap both forearms in paper towel, and cover those with plastic wrap to prevent evaporation. (wash/dry forearms before this.)
  • perform activity until paper towel saturated.
  • collect sweat in small plastic bag, as this is a good shape for allowing the meter to operate with a small quantity of liquid.

I’ve done a ‘dry run’ of this skipping some of the steps required for accuracy, and have a number that looks reasonable. My sweat measured 4200 ppm NaCl, so using the formula below and adjusting just for Sodium this gives me: (4200 * .57 + .618)0.3930.8 = 752 mg/liter of Sodium in my sweat. (0.393 is Sodium proportion of Na in NaCl, 0.8 is sweat being 80% sodium.) This is a lower number than I was expecting, as I have always seemed to be a salty sweater.

I plan on soon doing a more controlled test with rinsed paper towels and arms.

Some open issues that I am still working on figuring out that I’d appreciate pointers on:

  • What are the negative ions in sweat? Ie is there typically a chloride ion for every sodium ion? What balances the charge in sweat? My initial searches didn’t show much on this.
  • How does the above answer and the 20% potassium affect how the TDS measurement works? This is really a matter of calibration, as for instance the Precision Hydration test is described as using electrical current to measure concentration as well.

Even with the above areas of uncertainty, my guess is that this is still going to give me a number with +/-20% accuracy. I think even a number with this type of accuracy can be helpful, as at least my initial numbers are telling me I am less than 1gm/Na per liter.

2 Likes

I’m interested in this scientific venture of yours. Good luck, looking forward to your full test. :+1:

1 Like

As a scientist, I admire your optimism… I think your error bars are going to be more like +/-50% and possibly more. Good luck!

i think you might be down a rabbit hole here but i will give it a go.

whats the symptoms here in terms of a problem? if you a cramping regularly then using any electrolyte in training and increasing the dosage might show improvements.

i have been looking into the prevision hydration as i appear to be really suffering from the key issues and essentially bringing myself close to hospitalisation every ironman i do. its all going in but i am missing the key elements for processing so i end up bloated, passing out worrying everyone.

training wise in lock down i jave been upping the high5 zero tabs in 750 water bottle hence doubling / tripling the concentration and i have been doing some serious long rides in heat.

been getting the same salty bibs but none of the night restless legs or cramps.

i think your testing whilst commendable is open to flaws, collected sweat will evaporate quickly so be more concentrated.

Using cling film you might artificially increase output of other products as the skin wont breath the same way as in open air so contaminate more.

i know gatorade have testing strips which you send off as well.

just thinking there might be a more obvious answer

Hi @toyman ,

Thanks for including us in the conversation here, it’s awesome to see you’re taking such an interest in your specific sweat losses and using some solid ingenuity to do some at-home testing.

Firstly, I’ll try and answer your two questions then I’ll add some thoughts on what you’re doing.

  • Negative ions in sweat include chloride, bicarbonate, sulfate and phosphate - though the latter three appear in trace amounts so more often than not go undiscussed. There’s a chloride ion for every sodium ion - sodium chloride (NaCl) is a typical ionic compound and each sodium ion is surrounded by six chloride ions and each chloride ion is surrounded by six sodium ions. The equal number of sodium and chloride ions balances the charge - there are other electrolytes (charged particles) and trace minerals in sweat but sophisticated, laboratory-based analytical techniques can exclude these (e.g. chromatography (IC), mass spectrometry, ion-selective electrode (ISE) and flame photometry (FP)). With the absence of eliminating the ‘minor’ electrolytes it may be best for you to simply accept that they’re there but largely insignificant (if you haven’t read it yet we’ve got a great blog talking about minor electrolytes and whether you need to consider them in the big picture, check it out here).

  • Unfortunately there’s no good answer here, when you’re measuring potassium content too you’ll always end up with a greater number than if you were measuring just sodium (at least that’s the assumption based on there being a greater amount of dissolved solids total). From what we’ve seen you’d be better off with something like a Horiba kit but they’re pretty expensive and from some testing we’ve done we’ve had variable results and certainly wouldn’t be 100% confident.

Now, some thoughts and comments on the whole process. I’m honestly stoked to see such an effort being made to understand individual needs in sweat losses - it’s literally what got us into this niche area years ago and we’ve done plenty of our own ‘backyard science’ ever since.

A few further things to consider as you continue testing your protocol:

Cleaning of the skin with deionised water prior to collection to avoid additional contamination.

Sweat should ideally be collected at the onset of sweating, not the onset of exercise, as sweat collected at the onset of exercise includes some skin surface contamination from residual sweat in ductal lumen (why we always wipe away the first bit of sweat when we test!). The onset of sweating can take between 20-90 minutes depending on an individual and their sweat rate; you want them to have reached steady state sweating so to confirm any contaminants have been flushed out.

Lastly, @stuart_steele1 is right, ‘patch testing’, especially that covered by clingfilm will create a microenvironment which through increasing local skin temperature and humidity will alter regional sweat rate compared with uncovered skin and hence may impact sweat sodium concentration

The biggest factor we’ve seen when looking at at-home testing of this sort is huge variation in results. It’s incredibly difficult to control all of the necessary variables, eliminate any contamination, and ultimately validate your results. And once you start getting varying results you then have no idea which ones to trust, or if you can truly trust any of them - that’s ultimately the big question in all sport science!

Now, I know you mentioned the price of an Advanced Sweat Test being prohibitive but should you want confirmation of your Sweat Sodium Concentration do please drop us a line on hello@precisionhydration.com and we can lend a helping hand getting you an appointment at your nearest Test Centre, once lockdown/shelter in place measures are lifted obviously!

Feel free to keep us in the loop with your continued testing and reach out if there’s anything else we can help with.

6 Likes

Thanks @PrecisionHydration for the response! I had planned to wash my forearms beforehand, so the tip regarding onset of sweating is a key point I had missed. I did find a table that gives the “TDS factor” for NaCl and KCl solutions of different concentrations, so that will help with adjusting the readings from my TDS meter that is calibrated for NaCl. For these conversions I’m planning to use 80 %NaCl and 20% KCl as the assumed sweat proportions. I know other stuff is there, but not much, and I have no way of using that to usefully adjust the TDS readings.

I think I’m ready to give the test a try, as I think I have a plan for controlling the variables that I am aware of. I guess I’ll need to do it several times, to see if at least my results are repeatable. While that doesn’t confirm their accuracy, it does at least give me an idea of how much noise I am getting in the data.
This all started as a ‘quick and dirty’ sweat test, but has slowly gotten more involved. I’m not thinking that I’ll end up getting a proper test to check my results, so there go my cost savings :slight_smile: I did look at your seat testing locations, and the closest are several hours away (I’m east of Sacramento). I am sometimes down in the San Jose area, so I my try to go to the Santa Cruz location at some point.

I’ll post updates when I make some progress.

No problem @toyman, we love to talk hydration and science!

We’d be keen to stay in the loop whilst you’re running your at-home testing and act as a soundboard for any hurdles you come across - as you’ve highlighted, repeatability is key, then you need to figure out the accuracy of your repeatable data!

I know Dave down in Santa Cruz would love to hear about your DIY testing and help provide some clarity with a PH Sweat Test.

Let us know how you get on with your at-home testing and if you do make it to Santa Cruz drop us a message after the test with your result!

1 Like

OK, some hot weather prompted me to get going on this, and I have my first results.
Conditions in my garage were 91F and 37% humidity for the test.

I collected sweat from both forearms, and measured the TDS (total dissolved solids) from each. This is measured in PPM NaCl:
Left: 2580
Right: 2302
I then mixed the collected sweat, and measured 2440 PPM. I will use 2440 for the calculations.

Using the formula (modified for units) from the Gatorade sports science web page to convert forearm sweat composition to whole body composition, I get the following:
Whole Body mg/L Na+ only = (0.57 * (forearm sweat NaCl PPM) + 618)*.393
This works out to 789 mg of Na/L whole body sweat. (959 mg/L from forearms.)

This seems to be right in the middle of the bell curve for saltiness, and I thought I was more salty than that :slight_smile: The Precision Hydration online sweat test puts my electrolyte losses as ‘high’.
I will also do some sweat rate tests. I did one concurrently with the above tests, but since I was off the bike a lot it’s hard to say what the conditions for that test were. I sweat about 2.25 liters over 2 hours, 1 hour of which was riding in the hot garage, the rest was inside at about 76F.

I’m pretty happy that the left/right arms came out relatively close. I plan to do some repeat tests in different conditions and see if I get repeatable results.

Some things I have done to try to minimize sources of error:

  • Wash arms after onset of sweating (Thanks @PrecisionHydration for the tip!) I rode for 15 minutes, then washed/rinsed my forearms. Final rinse in reverse osmosis(RO) water then dried them before wrapping them to collect the sweat.
  • Added RO water to some plastic bags to check to see if either the bag, the plastic wrap, or the rinsed/dried paper towels would add measureable solids. The plastics didn’t add anything, and the paper towel only added about 6, so I’m ignoring that for now. (I guess I should have subtracted 6…)
  • the paper towels I am using are soaked/rinsed in RO water, then dried. Without this step they add significant solids to the water.
  • Checked the TDS meter with a calibration solution - it measured within 2.5% of the calibration solution. I did not adjust for this, but it measured the standard at 973 instead of 1000, so if I scaled the results for this, it would increase the salt content slightly.
  • Checked how the TDS meter measures NaCl vs KCl. At the relative concentrations, the ‘TDS factor’, or the conversion from the measured conductivity of the water to the TDS value is within a few percent, so I am not trying to correct for this, but I do not think it is a significant source of error.
2 Likes

I am glad you tried it at least, necessity is the mother of all invention

Hi @toyman,

Thanks for the update.

Firstly, great job, it sounds like your initial testing has gone really well and it’s great to hear that you took some of our advice on board.

As you say, retesting as much as possible over the next couple of weeks is really important. When you do this, it would be optimal if you try and keep as many of the variable the same as possible (i.e. same training session, same temperature, same kit etc.) to try and replicate the same sweat rate.

As for the result itself, it’s very plausible. Although, anecdotally you’ve suggested that you anticipated a higher result which always raises questions as we know that most athletes who think they’re ‘salty’ generally are.

Once you’ve pooled some repeat data and feel like you’ve landed on a score that you’re confident is at least valid via your at-home protocol we’d be happy to discuss the implications in my detail over email, simply reach out via hello@precisionhydration.com

OK, so I’ll kick my own old thread, as I finally got around to doing some more testing, and thought I’d provide an update. (previous posts go over how I have been doing this.)
@Dr_Alex_Harrison - I thought this might interest you.

I’ve now done 3 tests - 2 a while ago, and one just recently.
The results in mg/L of Sodium for the tests are:
800mg/L (5/2020) 92F
920mg/L (7/2020) 87F
1148mg/L (9/2022) 94F
Note that these are whole body sweat concentration calculated from the forearm sweat concentration using equation 3 from: https://www.gssiweb.org/sports-science-exchange/article/sse-161-sweat-testing-methodology-in-the-field-challenges-and-best-practices

As @PrecisionHydration mentioned, there is quite a bit of variability here, so I haven’t ended up with a ‘precise’ number to use. Also I found that my left arm always produced higher concentration, by 12, 16 and 24% respectively, so as the overall concentration went up the left arm difference went up as well. I think I have controlled for variability about as well as is practical - washing arms with RO water, waiting for onset of sweating, calibrating TDS meter with calibration solution among other things. This has grown to no longer be a ‘simple DIY’ solution - based on the results I can’t recommend anyone trying this. The main reason to do something like this is to practice running experiments, not to get useful data :slight_smile:
At the end of the day, my assumption going into the testing was that my sweat concentration was likely about 1g/L based on surveys, etc, and the test results are consistent with that.
I also measured a sweat rate of about 1.6 L/hour on my trainer at 94F, so during the 2 hour test ride I did I lost about 4 grams of Sodium during the test, or almost 11 grams of NaCl. Even though these numbers are approximate, I do think it is helpful to roughly quantify these to guide replenishment. If I were to get all this from some 1g NaCl tablets I have, this would have been 1 tablet every 10-15 minutes on the bike, which surprised me.

Couldn’t agree more. There is a better way!

Read the literature as you did. Make a guess. Aim on the slightly high end for sodium replacement. Be quantified in your approach, but listen to and observe your body’s own feedback. No training interruption needed. Far more time efficient, as you noted!

For those reading:
If drinking lots, peeing lots, and losing weight → intake more sodium.
If drinking lots, not peeing, and losing weight → intake more fluid and more sodium in direct proportion to more fluid.
If peeing lots and maintaining weight → drink less fluid and take more sodium.
If not peeing and maintaining weight → probably in the ballpark, but maybe more fluid is better, if activity duration is >4 hrs. Need to urinate is sign of good hydration and necessary in long activities.

If HR drifts up for power and not peeing → more fluid.
If HR drifts up for power and peeing lots → more sodium.
If HR does not drift up for power and peeing lots → less fluid.
If HR does not drift up for power and not peeing → you’re probably in the ballpark.

If thirsty and peeing lots → intake more sodium and less fluid.
If thirsty and peeing little → more fluid and more sodium.
If not thirsty and peeing lots → more sodium.
If not thirsty and peeing little → might be in the ballpark but might also be way low sodium and need a bit more to drive thirst up. Check HR drift. If drifting, more sodium and more fluid.

If cramping → consider increased sodium and maybe increased fluid but if increasing fluid, increase sodium by more.
If GI distress → consider increased fluid, or lower total consumption of carbs & sodium, or better gluc:fruc ratio.

Usually the result of sweat testing is:

  1. Overconfidence in inaccurate sodium intake needs.
  2. Ignoring what the body is already telling us. (urination, body weight loss, thirst, cramping, fatigue, cardiac drift, etc, all tell us more than enough to make very good guesses about what needs to happen)
  3. High time cost.
  4. Confirmation of what the literature tends to tell us.
  5. Forgetting that sweat sodium concentrations and sweat rates change dramatically with changing conditions, fitness, training experience, body weight changes, clothing choice, absolute intensity, relative intensity, etc. Ignoring those things leads to rigidly following a plan that was designed for different conditions.

In your case, you tested often enough to notice that “stuff changes” and “results are variable.” Good on you!

Yep. Salt tablets have a ridiculously low amount of sodium in them. Most folks should intake 500-1500mg per hour. If sweating heavily, definitely target the top end. Don’t exceed 1500mg per liter of water…mayyyybe 1800mg/L. Depending on the salt tab, you’d have to take a salt tab every 10-15 minutes for the entire duration of the activity, LOL. Maybe twice as often, in the case of some ridiculous products.

2 Likes