The only carbohydrate fractions that matter are those that increase insulin.
Nonstructural Carbohydrates (NSC), Non Fiber Carbohydrates (NFC), Ethanol Soluble Carbohydrates (ESC), Water Soluble Carbohydrates (WSC) - it all really comes down to hydrolyzable carbohydrates for metabolically challenged horses.
Hydrolyzable carbohydrates (HC) are defined as those digestible in the small intestine. Some microbial fermentation occurs in the small intestine but much less than in the hindgut, so hydrolyzable carbohydrates will primarily be absorbed intact.
Starch is digested entirely to glucose. Simple monosaccharide sugars like glucose and fructose are absorbed intact. Disaccharide sugars like sucrose (table sugar) and lactose (milk sugar), and the trisaccharide trehalose, are broken down into monosaccharides by enzymes present at the brush border of the small intestinal cells, then absorbed. These are the HC and the only components that increase blood glucose and therefore insulin.
More complex plant sugars like cellobiose and raffinose cannot be digested and pass through to the hindgut where they are fermented to volatile fatty acids. The same fate awaits all fiber fractions, "resistant" starch not accessible to the digestive enzymes and rapidly fermentable carbohydrates like pectin, fructan and beta-glucan.
The volatile fatty acids/fermentation products produced can substitute for glucose in energy pathways (acetate), be used to feed the colonic cells (butyrate), converted back to glucose or glycogen (propionate, lactate) or fat (butyrate, acetate). None of those substances or reactions will increase insulin. In metabolic horses, it is insulin that causes laminitis.
Forage analysis is geared toward ruminants rather than the simple-stomached horse, so the distinction between HC and fermented doesn't mean as much because they both get fermented in the forestomachs of a ruminant. For now, we have the most data on sugars in hydrolyzable carbohydrates estimated colorimetrically by Ethanol Soluble Carbohydrates (ESC). The sum of ESC + starch will give you the best estimate of total HC in a forage sample – not NSC.
Of 264 independently collected, good-quality grass hay samples on file, 97% are under 10% ESC + starch, but only 15% are under 10% NSC. If using NSC as a threshold, then owners may opt for low to poor quality hays in an effort to achieve that threshold. Even if the threshold is elevated to an arbitrary 12%, only 35% of forage in this database would meet that criterion. Using NSC as a measure of safety places an unnecessary burden on owners of horses with metabolic disorders.