Technical Note: A method for determination of titanium dioxide concentration in fecal samples

Abstract

Titanium dioxide has been used as a marker for determining diet digestibility indirectly, but some authors have expressed difficulty in measuring TiO₂ concentrations in fecal material. We developed an accurate and precise method to determine TiO₂ concentrations in equine feces. The method includes dry-ashing samples, digestion with (NH₄)₂SO₄ in concentrated sulfuric acid, followed by the addition of H₂O₂ to produce a yellow to orange color that can be read spectrophotometrically. Accuracy was tested by spike recovery, and precision was tested by examining the coefficient of variation (CV) between duplicates of 449 individual samples. The method described here was compared with a previously published method by examining CV between duplicates of samples analyzed using both methods and comparing them using a paired t-test. Titanium dioxide spike recovery averaged 106%, and the CV between duplicates averaged 4.0%, with 79% of sample pairs having a CV of <5%. When compared with a previously published method, the method described here had a lower CV between duplicates (P < 0.0001). The method described here provides an accurate and precise quantitative analytical procedure for TiO₂ in equine fecal samples.

Lay Summary

Titanium dioxide is a marker fed to animals to help determine diet digestibility indirectly by measuring the concentration of TiO₂ in fecal samples. This paper describes an accurate and precise method to analyze TiO₂ in equine feces. The precision of this method is demonstrated by the low variation between sample duplicates. This method requires fewer sample replicates for analysis, leading to less labor, expense, and waste in the laboratory.

A method for analyzing equine fecal samples for TiO₂ is described. Measuring TiO₂ in feces may be useful for measuring diet digestibility indirectly.

Introduction

Determining diet digestibility using the total fecal collection method is time consuming and labor intensive, and requires that animals wear collection harnesses or be housed in facilities that allow separation of urine and feces. The use of indigestible markers to estimate diet digestibility offers advantages in terms of animal management and ease of sample collection.

Titanium dioxide has been shown to be useful as an external marker for digestibility studies in many species and to perform just as well or better than Cr₂O₃ (Short et al., 1996; Titgemeyer et al., 2001; Myers et al., 2006). Titanium dioxide is favored over Cr₂O₃ because AAFCO lists TiO₂, but not Cr₂O₃, as a legal additive for animal diets (AAFCO, 2015), but some authors have expressed difficulty in measuring TiO₂ concentrations in fecal material (Titgemeyer et al., 2001; Myers et al., 2004).

One method for determining TiO₂ concentrations in digesta or feces was developed by Short et al. (1996) and involves dry-ashing samples, digesting in acid, and inducing a colorimetric response of TiO₂ by adding H₂O₂. A modification to this method by Titgemeyer et al. (2001) has been reported. Myers et al. (2004) developed a wet-ashing procedure for determining TiO₂ concentrations in feces and used K₂SO₄ and CuSO₄ as reaction catalysts to facilitate the dissolution of TiO₂. However, our lab has had difficulties producing consistent and accurate results for TiO₂ concentrations in duplicate fecal samples using the methods of Short et al. (1996) as modified by Titgemeyer et al. (2001). The main difficulty that our laboratory has encountered with the method is incomplete dissolution of TiO₂ during digestion. Varying amounts of dissolution lead to large variation among replicates. Often, more than two replicates per sample are needed to achieve an acceptable coefficient of variation (CV) between replicates (Adedokun et al., 2012; Rochell et al., 2012; Morgan et al., 2014), which increases expense and slows the analytical process. Therefore, the objective was to develop an accurate and precise method for the determination of TiO₂ concentrations in feces using duplicate samples.

Materials and Methods

Practices and protocols associated with animals were approved by the Institutional Animal Care and Use Committee at the University of Kentucky.

This procedure involves dry-ashing fecal samples, digesting with (NH₄)₂SO₄ in concentrated sulfuric acid, followed by the addition of H₂O₂ to produce a yellow to orange color, and measuring this solution spectrophotometrically at 410 nm (Table 1). The main change from the procedure described by Short et al. (1996) as modified by Titgemeyer et al. (2001) was the addition of (NH₄)₂SO₄ during digestion.

Table 1.

Procedure for determining TiO₂ concentration in fecal samples

Step	Instruction
Standard preparation
1	Dissolve 250 mg TiO2 in 100 mL of concentrated H2SO4 with heating and stirring.
2	Once dissolved, add to a 500-mL volumetric flask containing ~200 mL deionized water. Add another 100 mL of concentrated H2SO4 and dilute to volume with deionized water. This is the stock solution for standards.
3	Prepare standards by pipetting 0, 0.5, 1.0, 1.5, and 2.0 mL of stock solution into 50 mL volumetric flasks. Add 10, 9.5, 9.0, 8.5, and 8.0 mL concentrated H2SO4 to each flask, respectively, so the combined volume is 10 mL. Add 10 mL of 30% H2O2 to each flask and dilute to volume with deionized water.
4	Measure absorbance of standards at 410 nm.
Procedure
1	Weigh 0.15 g of sample into crucibles and heat in ash oven at 600 °C for at least 8 h.
2	Rinse ash from cooled crucible into 250 mL digestion tubes using deionized water.
3	Add 1 g (NH4)2SO4 and 13 mL of concentrated H2SO4 to each tube.
4	Heat tubes in digestion block at 420 °C for 3 h.
5	Rinse contents of digestion tube into 50 mL volumetric flasks containing 10 mL of 30% H2O2 and dilute to volume with deionized water.
6	Let flasks sit overnight and measure absorbance of solution at 410 nm.

Standard preparation

Standards were prepared from a stock solution with a concentration of 0.5 mg TiO₂/mL. The stock solution was prepared by dissolving 250 mg TiO₂ in 100 mL concentrated H₂SO₄ with heating (~250 °C) and stirring on a hot plate. Once dissolved and cooled, this solution was added to a 500-mL volumetric flask containing approximately 200 mL deionized water. Another 100 mL of concentrated H₂SO₄ was added and then the solution was diluted to volume (500 mL) with deionized water. Standards (0, 0.005, 0.010, 0.015, and 0.020 mg TiO₂/mL) were prepared by pipetting 0, 0.5, 1.0, 1.5, and 2.0 mL of the stock solution into 50 mL volumetric flasks and then adding concentrated H₂SO₄ to each flask so that the combined volume was 10 mL (10, 9.5, 9.0, 8.5, and 8.0 mL concentrated H₂SO₄ for each standard, respectively). To each flask, 10 mL of 30% H₂O₂ was added and the solution was diluted to volume (50 mL) with deionized water. The absorbance of standards was read at 410 nm using the 0 mg TiO₂/mL standard as the blank. Standards were run with every set of digested samples.

Fecal samples

Fecal samples (n = 449) were obtained from 15 horses that were fed diets varying in forage content. Five horses were fed a high forage diet (95% timothy hay, 5% supplement pellet; dry matter intake [DMI] = 1.2% of body weight [BW]), five horses were fed a low forage diet (45% timothy hay, 55% concentrate; DMI = 1.4% of BW), and the remaining five horses were fed an intermediate forage diet (73% timothy hay, 27% concentrate; DMI = 1.4% of BW). A combination of chopped and cubed timothy hay was used, and all ingredients for each horse were mixed to form a total mixed ration.

Titanium dioxide was fed at a rate of 2.00 g TiO₂/kg DMI per day for 3 to 4 wk before fecal collections. The daily ration and daily TiO₂ dose were equally divided into three meals fed every 8 h for at least 12 d before fecal collections. The TiO₂ dose was thoroughly mixed into each meal.

Fecal samples were obtained using fecal collection harnesses that were placed on the horses for a 5-d collection period. Throughout the 5-d period, fecal samples (250 g) were removed from the bag approximately every 4 h (30 samples collected per horse). Samples were immediately frozen until they could be dried in a 55 °C forced-air oven and ground to pass through a 2-mm screen. At one collection point, one fecal sample was not collected for one horse fed the low forage diet, resulting in 149 samples for the low forage diet compared with 150 fecal samples collected for the other two diets.

Procedure

Expected fecal TiO₂ concentrations were calculated by dividing TiO₂ intake (g) by expected fecal output (kg). Based on the assumption that TiO₂ is not absorbed by the horse and that an estimation of fecal output is based on expected diet digestibility, the expected concentration of TiO₂ in the fecal samples was calculated to be between 0.3% and 0.7%. Appropriate fecal sample sizes that would yield results within the range of the standard curve were calculated to be 0.04 to 0.33 g. In the literature, authors reports using various sample sizes, ranging from 0.1 (Short et al., 1996) to 10 g (Leone, 1973), likely dependent on method and amount of expected TiO₂. We decided to use a fecal sample weight of 0.15 g, as the expected TiO₂ concentration would fall in the middle of the standard curve.

All glassware was washed with mineral-free, acidic detergent (Contrex CA Acidic Liquid Detergent, Decon Labs, Inc., King of Prussia, PA) and rinsed with distilled water before each use. Duplicate dry and ground fecal samples (0.15 g) were weighed into quartz crucibles and ashed overnight, or at least 8 h, at 600 °C in a muffle furnace. Ash was then quantitatively rinsed into 250 mL digestion tubes (Foss, Hillerod, Denmark) using distilled, deionized water. One gram of (NH₄)₂SO₄ was added to each digestion tube, followed by the addition of 13 mL of concentrated H₂SO₄. Digestion tubes were placed in a digestion block with a tube rack (FOSS Digestor 2520, Foss, Hillerod, Denmark). Samples were digested at 420 °C for 3 h. After digestion, contents of the digestion tubes were quantitatively rinsed using distilled, deionized water into 50 mL volumetric flasks containing 10 mL of 30% H₂O₂. Flasks were then diluted to volume with distilled, deionized water and carefully mixed. Instead of filtering the solution as others have done (Short et al., 1996; Myers et al., 2004), flasks were allowed to sit overnight so that any acid-insoluble ash would settle to the bottom. An aliquot of the supernatant was transferred to a disposable 3 mL cuvette, and absorbance of the solution was read at 410 nm (Genesys 10S UV-Vis; Thermo Fisher Scientific, Waltham, MA, USA). The TiO₂ content of the samples was calculated using the equation resulting from the standard curve that plots the amount of TiO₂ in each standard against absorbance.

Procedure validation

Percent recovery of TiO₂ from spiked samples was used to assess the accuracy of this method. Three different fecal samples, with previously analyzed known concentrations of TiO₂, were spiked with 2.0 to 4.5 mg of TiO₂. Percent recovery of TiO₂ was calculated based on the amount of TiO₂ spike added to the sample and the amount of TiO₂ analyzed (adjusted for the TiO₂ already present in the sample, as measured previously). Regression was used to evaluate the relationship between the amount of TiO₂ spiked in the sample and the amount of TiO₂ recovered from the spiked samples. A paired t-test for equivalence was used to determine if the actual and analyzed amounts of added TiO₂ in spiked fecal samples were similar (SAS 9.4, SAS Institute Inc., Cary, NC).

The CV was calculated for duplicate samples and used to assess method precision. The CVs were averaged across all diets and were compared using an analysis of variance (ANOVA; SAS 9.4, SAS Institute Inc., Cary, NC). The frequency of CV lower than 5%, 10%, and 15% was calculated for each treatment and compared using a Chi-square test (SAS 9.4, SAS Institute Inc., Cary, NC).

Method comparison

Coefficients of variation between duplicates were also used to compare the modified method described here to a previously published method. Twenty-three duplicate fecal samples, with expected TiO₂ concentrations ranging from 0.25% to 0.50%, were digested and analyzed using the method described here as well as the dry-ash method described by Short et al. (1996) as modified by Titgemeyer et al. (2001). The CV between duplicates was calculated for each sample pair, and the mean CV was compared between methods using a paired t-test (SAS 9.4, SAS Institute Inc., Cary, NC). Regression was used to assess the relationship between the concentrations of TiO₂ when analyzed using each method.

Results and Discussion

As the amount of TiO₂ added to the sample increased, the amount of TiO₂ recovered from the added TiO₂ spike also increased (Figure 1). The slope of the regression line (slope = 1.01) was close to 1, and the y-intercept was close to zero (intercept = 0.17), indicating that this procedure was accurate in determining TiO₂ concentrations in fecal samples. Percent recovery of the added TiO₂ from spiked samples averaged 106%, and the amount of TiO₂ added as the spike was the same as the amount of recovered TiO₂ (equivalence test; P < 0.0001).

Figure 1.

Relationship between increasing amounts of TiO₂ added to samples as a spike and the amount of TiO₂ recovered from the spike.

Method precision is also an important characteristic of procedure development and was assessed in this experiment by analyzing variation between duplicate samples. For all 449 equine fecal samples analyzed, the average CV was 4.01 ± 0.29%. When broken down by diet, the high forage diet had a larger mean CV than the low or intermediate forage diets (P < 0.05; Table 2). However, all diets had a mean CV less than the recommended maximum CV of 15% for bioanalytical assays (Tiwari and Tiwari, 2010). Seventy-nine percent of all samples had CV less than 5% and 97% of samples had CV less than 15%. Only 13 samples out of the total 449 samples (low forage diet = 2 samples, intermediate forage diet = 2 samples, and high forage diet = 9 samples) would need to be rerun if using the recommended maximum CV of 15%.

Table 2.

Mean (±SE) coefficient of variation (CV) between duplicates and percentage of samples with CV less than 15%, 10%, and 5% for three diets varying in forage content

Diet1	Mean CV ± SE (%)	Percent of samples with CV < 15% (%)	Percent of samples with CV< 10% (%)	Percent of samples with CV< 5% (%)
High forage	5.06 ± 0.71a	94y	89	76
Intermediate forage	3.49 ± 0.34b	99x	95	81
Low forage	3.48 ± 0.38b	99x	95	79

High forage diet contained 95% timothy hay and 5% vitamin–mineral pellet; intermediate forage diet contained 73% timothy hay, 14% beet pulp, 6% whole oats, 2% soybean oil, and 5% vitamin–mineral pellet; and low forage diet contained 45% timothy hay, 25% beet pulp, 16% whole oats, 9% soybean oil, and 4% vitamin–mineral pellet.

Means within a column lacking a common letter differ (P < 0.05). Analysis of variance (ANOVA) was used.

Means within a column lacking a common letter differ (P < 0.05). Chi-square test was used.

When comparing among diets, the low and intermediate forage diets had a greater percentage of samples with CV less than 15% compared with the high forage diet (P < 0.05; Table 2), meaning that more samples would need to be rerun from the high forage diet. When using a more stringent CV cutoff value (i.e., 10% or 5%), there were no longer differences among diets (P > 0.05; Table 2). It is possible that the fecal sample size (0.15 g) used in this method may have been too small to be representative for a horse fed a high forage diet, leading to larger CV between duplicates. Perhaps a larger fecal sample size would improve method precision for feces from horses fed high forage diets. Understanding how diet impacts method precision is important to consider when planning and budgeting for sample analysis.

This method was compared with a previously published method (Short et al., 1996; Titgemeyer et al., 2001) by comparing CV between duplicates for samples analyzed using each method. Under our laboratory conditions, the mean CV between duplicates for the method described by Short et al. (1996) as modified by Titgemeyer et al. (2001) was larger than the mean CV between duplicates for the current method described in this study (Figure 2; P < 0.0001). Only 26% of the samples analyzed using the previously published method had CV less than 15%, while 100% of the samples analyzed using the method described here had CV less than 15%. Additionally, the relationship between the concentrations of TiO₂ when analyzed using both methods was not strong (Figure 3; P > 0.10). The method described here did not produce consistently greater or consistently lower concentrations compared with the previously published method, but it did produce more precise results as indicated by the lower CV between duplicates.

Figure 2.

Mean coefficient of variation (CV) between duplicates for samples analyzed using the dry-ash method described by Short et al. (1996) as modified by Titgemeyer et al. (2001) and the method described here (n = 23). Means differ (P < 0.0001).

Figure 3.

Titanium dioxide concentration in the same samples when analyzed by the method described here compared with the dry-ash method described by Short et al. (1996) as modified by Titgemeyer et al. (2001). The relationship between the two variables is poor (P > 0.10; R² = 0.0008). The dotted line indicates a perfect relationship (R² = 1.0).

This method produced consistent, reproducible results without the need to utilize more than two replicates per sample. Other papers report that triplicate (Adedokun et al., 2012) or quadruplicate (Rochell et al., 2012) replicates were used to analyze for TiO₂ using other methods, suggesting that duplicate replicates were insufficient to obtain precise results. The use of duplicate samples, instead of triplicate or quadruplicate samples, not only reduces the total cost of analysis and time required to analyze samples, but it also reduces the generation of lab waste, which has become increasingly important.

Implications

The method described in this paper provides an accurate and precise quantitative analytical procedure for TiO₂ determination in fecal samples. The ability of this method to determine TiO₂ concentration in equine feces, in particular, encourages the testing of TiO₂ for use as an indigestible marker in horse digestion studies, which would reduce expense, labor, and waste in performing digestibility studies with horses.

Glossary

Abbreviations

BW

body weight

DMI

dry matter intake

Conflict of interest statement

The authors have no conflicts of interest to declare.