Don't be Fooled - Malic Acid Testing

I recently attended the Unified Wine & Grape Symposium in Sacramento.  A few winemakers I spoke with asked about the precision of the Accuvin Malic Acid test kit.  As has been reported, the accuracy of the Malic Acid test kit is +/- 10 mg/L at the 30 mg/L level, which is generally considered the end of MLF, and about +/- 20 mg/L at the 75 mg/L level.

These questions reminded me of the of the Wine Industry Interlaboratory Testing Program that has be going on since 1999.^1,2,3  The program, set up by the ASEV and UC Davis, is meant to help winery labs improve the quality of their results.  In this program, two wine samples are sent out a few times a year to the 40 – 55 participating winery labs.  The samples are analyzed at least in duplicate, and the results are sent back for statistical analysis.  The most common method used by far was enzymatic.  A table of the results is given below for wine samples with malic acid levels below 500 mg/L.  From 10% to 23% of the results were found to be unreliably inaccurate, and are not included in the table.

What do these results show?

First, in the nine years the quality improvement program has been going on, the results have gotten progressively worse.  The coefficient of variation, a measure of lab to lab reproducibility, has doubled.  In addition, while in the initial years 10% of the reported results were so far off the mark that they had to be excluded, the number of totally unusable results climbed to 23% in the latest report.

Second, the results have deteriorated to the point where about half are flat out wrong regarding what they tell the winemaker.  In 1999, a sample with a malic acid level of 100 mg/L was analyzed, with the results showing a range of 63 – 137 mg/L.  That’s pretty broad.  In 2007 a wine with a malic acid level of 110 mg/L was analyzed, and the results ranged from 30 – 200 mg/L.  In 2008 a sample with a malic acid level of 96 mg/L was analyzed, and the range for the results had ballooned to 18 – 178 mg/L.  Some winemakers were told their wines had finished MLF (which they hadn’t), while other winemakers thought they had a few weeks yet to go (which they didn’t).  And remember, that huge range was the range reported AFTER 23% of the results had been discarded for gross unreliability.  (By the way, if an Accuvin test strip had been used to analyze the above samples, you could expect the results to be in the range of 75 – 125 mg/L.  That’s considerably better!)

In 2002 Dr. Christian Butzke of UC Davis wrote a critique of the first few years of the program.  Regarding malic acid he wrote:  “Given the corresponding consequences of uncertainty over incomplete or deliberately prevented malolactic fermentation, in particular CO2 evolution in the bottle and changes in perceived acidity, this performance requires major improvements.”¹ What would he say today?

Don’t be fooled!  Spending money on expensive equipment or excessive time running laboratory tests does not ensure reliable results.

A final thought:  just getting the same answer twice with a method performed unreliably does not make it the correct answer.

1   C. E. Butzke, Am. J. Enol. Vitic., 53 (20: 163 – 169 ) 2002.

2   C. E. Butzke, Practical  Winery & Vineyard, Jan. Feb. 2002

3   Collaborative Testing Services, Inc., www.cts-interlab.com/wine/index.html, reports, 2003 - 2008