Background: Cutoff activities for diagnosing disaccharidase deficiencies are historical and are difficult to verify from a reference population. The objectives of this study were to validate the utility of historical disaccharidase cutoffs using data from clinical samples and to evaluate the demographics of individuals for whom intestinal disaccharidase testing was performed.
Methods: Results from 14,827 disaccharidase test samples were extracted from the laboratory information system. Data were analyzed by the Hoffman method to calculate a reference interval for each enzyme, and the lower limits were compared to historical cutoffs. The observed frequencies of disaccharidase deficiencies were determined using historic and calculated cutoffs.
Results: The median patient age of the entire data set was 13 years (range <1–88 years), and 45% were male. The cutoffs for lactase, maltase, palatinase, and sucrase were determined to be 10, 100, 9, and 25 U/g protein, respectively. Applying these cutoffs to the data set, 61% had no enzyme deficiencies, 35% were lactase deficient, 11% were maltase deficient, 13% were palatinase deficient, and 13% were sucrase deficient. Pandisaccharidase deficiency was present in 8%.
Conclusions: Disaccharidase testing is most commonly performed in patients <18 years. Lactase deficiency is the most frequently observed single-disaccharidase deficiency. The historical cutoffs for maltase and sucrase were validated. To align with calculated reference intervals, the palatinase cutoff should increase from 5 to 9 U/g protein, and the lactase cutoff should decrease from 15 to 10 U/g protein.
A deficiency of one or more intestinal disaccharidase enzymes can result in carbohydrate maldigestion. The measurement of their activities in small bowel mucosa is considered to be the gold standard test for a deficiency diagnosis. Establishing reference intervals for disaccharidase activities is challenging because biopsy of the small bowel is invasive; thus, reference samples are not available for testing. We applied the Hoffman method to a large cohort of clinical results to validate historic deficiency cutoffs. Results indicate historic cutoffs for maltase and sucrase were valid, but changes are recommended for lactase and palatinase. Use of appropriate cutoffs is important to identify individuals with disaccharidase deficiency.
Dietary disaccharides are important sources of exogenous glucose. The small intestine is impermeable to disaccharides, requiring enzymatic hydrolysis of disaccharides into monosaccharides. This hydrolysis is catalyzed by disaccharidase enzymes, which are localized to the brush-border membrane of the small intestine. Decreased or absent activity of one or more of the disaccharidase enzymes can result in carbohydrate maldigestion, which is commonly characterized by abdominal symptoms including cramping, flatulence, and diarrhea (1, 2).
The main dietary disaccharides are lactose, maltose, and sucrose. The disaccharidase activities that are of clinical interest are lactase, sucrase, maltase, and isomaltase (more commonly referred to as palatinase). The disaccharidases responsible for hydrolysis of these sugars are the enzyme complexes lactase-phlorizin (EC 22.214.171.124), sucrase-isomaltase (EC 126.96.36.199), and maltase-glucoamylase (EC 188.8.131.52) (1). As enzyme complexes, they contain more than one active site and may display specificity for more than one substrate. Because of overlapping activities of the disaccharidase enzyme complexes, some measured activities, such as that of maltase, represent the aggregate activity of multiple enzyme complexes (maltase-glycoamlyase and sucrose-isomaltase). Additionally, deficiency of a single enzyme (sucrase-isomaltase) often results in decreased activities of maltase, sucrase, and palatinase.
Primary (hereditary) disaccharidase deficiencies are rare and include congenital sucrase-isomaltase deficiency, an autosomal recessive disease caused by a mutation in the sucrase-isomaltase gene; glucose-galactose malabsorption, due to a glucose transporter deficiency; and starch malabsorption, due to glucoamylase deficiency. Acquired deficiencies are much more common and are associated with injury to the intestinal mucosa. Such injuries may be due to anatomic anomalies or autoimmune disorders that affect the gastrointestinal tract. Deficiencies may also be transient due to extended antibiotic or other drug uses.
Establishing reference intervals for disaccharidase activity is challenging because these tests involve obtaining a biopsy of the small bowel mucosa and reference samples are not available for testing. Laboratories that perform disaccharidase testing have often adopted historical cutoff activities that are used to identify a disaccharidase deficiency (lactase, <15 U/g protein; maltase, <100 U/g protein; palatinase, <5 U/g protein; and sucrase, <25 U/g protein). The source of these cutoffs is not clear. The utilization of patient test results to establish a reference interval was originally described by Hoffman in 1963 (3) and has been shown to be accurate and reproducible (4, 5). The primary objective of this study was to retrospectively validate these historic cutoffs using the Hoffman method for reference interval determination. The secondary objective of this study was to evaluate the demographics and disaccharidase activities of individuals for whom disaccharidase testing was clinically performed.
The results of all samples for which disaccharidase activity testing was performed at ARUP Laboratories between November 2009 and May 2013 were extracted from the laboratory information system (LIS)3 and deidentified (n = 14,886 samples). Results were excluded if they were determined at a laboratory other than ARUP (n = 7), identified as an “LIS test” sample (n = 11), were associated with an absurd patient age of >110 years (n = 5), or failed to produce any reportable result (n = 36). The results from the remaining 14,827 samples were included in subsequent analyses. The University of Utah Institutional Review Board approved this project and its protocols.
Disaccharidase activity determination
The method used in our laboratory to determine disaccharidase activities has been described (6). Briefly, biopsies of small intestine mucosa were homogenized in saline and the homogenate incubated with each of four substrates (lactose, maltose, palatinose, and sucrose) to liberate glucose. After stopping the reactions by heat inactivation, the glucose concentration was determined using an automated chemistry analyzer. One unit (U) of disaccharidase activity was calculated as the quantity of enzyme that will hydrolyze 1 μmol disaccharide substrate per minute at 37 °C. Because of inherent variations in specimen size, the disaccharidase activity is determined relative to the protein concentration of the homogenate that was determined using a Lowry protein assay (7).
Determination of reference intervals
The reference intervals for disaccharidase activities were determined using the Hoffman method (3, 4). All enzyme activity results of 0 U/g protein were excluded from this analysis only (n = 408, 24, and 20 for lactase, palatinase, and sucrase, respectively). Enzyme activities were log-transformed before determining cumulative frequency distributions. Presumed outliers were not excluded from the data set because their influence on the cumulative distribution frequency was minimal and their exclusion did not make the analysis more scientifically or methodologically robust. The linear portion of the cumulative frequency distribution for each enzyme was visually estimated and used for linear regression analysis. The reference interval (RI) was calculated as RIlower = 2.5(m) + b and RIupper = 97.5(m) + b, where m is the slope and b is the y intercept of the linear regression. The calculated lower limit of the reference interval was used as the calculated cutoff activity. The cutoff ratio, defined as the ratio of the historical cutoff to the calculated cutoff, was determined for each enzyme. Microsoft Excel version 14.0.7166.5000 and GraphPad Prism version 5.0f software were used for all data analyses.
Disaccharidase reference intervals
The frequency of the log of enzyme activity for maltase, palatinase, and sucrase displayed unimodal distributions, while lactase showed a bimodal distribution (Fig. 1). This pattern was reflected in the cumulative frequency distributions of each enzyme, with an inflection point observed in the plot for log of lactase activity (Fig. 2). The linear portion of the cumulative frequency for each enzyme was used for linear regression analysis, and the upper and lower limits of the reference interval were calculated, as well as the cutoff ratio (Table 1). The cutoff ratios for maltase and sucrase were close to 1 (0.95 and 0.96, respectively), indicating close agreement between the historical and calculated cutoff values. The cutoff ratios for lactase (3) and palatinase (0.56) indicate discrepancies between the historical and calculated cutoffs, with the calculated lactase cutoff being 3 times lower than the historical cutoff and the calculated palatinase cutoff being almost 2 times higher than the historical cutoff.
The bimodal distribution of lactase results and the presence of an inflection point in the cumulative frequency of log activity (Fig. 2) suggested that two distinct patient populations were represented in the data set: a lactase-deficient and a lactase-sufficient population. The inflection point of the cumulative frequency distribution corresponds to a lactase activity of approximately 10 U/g protein, and this activity was used as a proposed cutoff to distinguish lactase sufficiency from deficiency.
Spearman correlations between disaccharidase activities greater than or equal to the modified cutoffs and patient age were −0.04 (P < 0.0001), −0.02 (P = 0.05), −0.08 (P = 0.01), and 0.007 (P = 0.42) for lactase, maltase, sucrase, and palatinase, respectively. On average, enzyme activities were approximately 5% lower in males compared to females.
Disaccharidase deficiencies by historical and proposed cutoffs
The disaccharidase activity results of all 14,827 patient samples in the data set were analyzed to determine the observed frequency of each disaccharidase deficiency, both as a single-enzyme deficiency and also in combination with other disaccharidase deficiencies, based on the historic and proposed cutoffs for enzyme deficiency (Table 2). The median patient age of the entire sample set was 13 years (range 0–88 years), and 82% of the samples were obtained from a pediatric (<18 years) population; 57.3% of the samples were collected from females. Using the historic cutoffs, 52.2% of the samples were not deficient in any disaccharidase, and this result increased significantly to 61.1% using the proposed cutoffs (P < 0.0001). The most commonly observed enzyme deficiency was lactase only, which was 34.6% using the historic cutoff. As expected, this decreased significantly to 23% with use of the proposed cutoff (P < 0.0001).
Overall, with or without additional enzyme deficiencies, lactase deficiency was most common using historic and proposed cutoffs (46.5% vs 34.8%, respectively). Likewise, using the historic cutoff, a palatinase deficiency was least common and occurred in only 3.2% of samples, but this increased to 13.3% using the proposed cutoff. This increase brought the frequency of palatinase deficiency close to that observed for sucrase and maltase (12.6% and 11.4%, respectively), which were unchanged by use of the proposed cutoffs.
A deficiency of all 4 enzymes (pandisaccharidase deficiency) was not common using the historic cutoffs (3.0%), but this result increased significantly to 7.9% using the proposed cutoffs (P < 0.0001). The increased prevalence of pandisaccharidase deficiency was identical to that reported in a separate study that also used a lactase cutoff of 10 U/g protein and a palatinase cutoff of 11 U/g protein to identify deficiencies (8).
Determining reference intervals is frequently challenging, especially when dealing with analytes measured in specimens that are difficult or invasive to obtain. Such is the case for the intestinal disaccharidases. However, the measurement of their activities is considered to be the definitive biochemical test for the diagnosis of disaccharidase deficiencies. The analytical method developed by Dahlqvist in 1964 (9) is widely considered to be the gold standard test and continues to be used, albeit with some technological modifications, by clinical laboratories. Interpreting the test results is aided by availability of well-defined reference intervals, yet these are impossible to obtain from a true reference population due to the invasive nature of the specimen collection. The source of the current cutoffs used by the majority of laboratories is not clear, and to our knowledge an investigation into the suitability of these historic cutoffs has not been conducted. The current analysis illustrates one approach for retrospective validation of reference intervals when it is difficult or impossible to obtain reference population samples.
The Hoffman method for determining reference intervals was retrospectively applied to data generated during clinical testing for disaccharidase deficiencies. The lower limits of the calculated reference intervals for lactase, maltase, palatinase, and sucrase were 5, 105, 9, and 26 U/g protein, respectively. The calculated lower reference limits for maltase and sucrase were consistent with the historic cutoffs of 100 and 25 U/g protein, respectively, and the cutoff ratios (historic/calculated) were 0.95 and 0.96, respectively. These data indicate that the calculated lower reference limit is consistent with the historic cutoff used to define enzyme deficiency. In contrast, the lactase cutoff was lower and the palatinase cutoff was higher than the historic cutoffs of 15 and 5 U/g protein, respectively. The cutoff ratio for lactase was a factor of 3, indicating the historic cutoff is greater than the calculated lower limit of the reference interval (5 U/g protein). The historic cutoff for lactase deficiency may be inappropriately high, resulting in over-diagnosis of lactase deficiency (46.5% of the subjects in this study). If a cutoff of 5 U/g is used as the criteria for diagnosing lactase deficiency, this would result in an observed rate of lactase deficiency of 20.5% in this data set. Based on the bimodal distribution of results, this cutoff is likely inappropriately low and would result in missed diagnoses of lactose deficiency. Based on the inflection point of the data (Fig. 2), the proposed lactase cutoff of 10 U/g would result in an observed rate of lactase deficiency of 34.8%. Interestingly, the reference interval for lactase activity was originally reported as 9–98 U/g protein (10) and, more recently, as 9–91 U/g protein (11). Both of these reference intervals correspond to the proposed lower limit for lactase activity, based on the observed bimodal distribution in this data set.
The cutoff ratio for palatinase was 0.56, indicating that the historic cutoff of 5 U/g protein is less than the calculated lower reference limit of 9 U/g protein. This finding suggests that the historic cutoff for palatinase deficiency may be inappropriately low, resulting in an under-diagnosis of palatinase deficiency. Importantly, because sucrase-isomaltase is an enzyme complex, it is unusual to encounter samples that are sufficient for one enzyme but deficient for the other. Using the historic palatinase cutoff, the frequency of all palatinase deficiencies was 3.2% and was unexpectedly lower than the frequency of all sucrase deficiencies (12.6%). Using the proposed palatinase cutoff increased the frequency of all palatinase deficiencies to 13.3%.
The disaccharidase activities were minimally correlated to patient age and were only slightly lower in males compared to females. While reference intervals partitioned by age and/or sex could be considered, the differences between them would be too small to be of clinical usefulness. A similar conclusion has been reached by others (11).
One limitation of the current study is that the population included was from a retrospective data analysis. While the Hoffmann method for reference interval determination is designed for use with a non-reference population, a general assumption is that the data set to which this technique is applied reflects the general population. A previous study evaluated the Hoffman method by applying it to well-characterized laboratory tests such as creatinine and thyroid-stimulating hormone (4). The author reported that applying their calculated reference limits for creatinine and thyroid-stimulating hormone to their study data set, the percentage of results that fell outside of those limits correlated with the prevalence of chronic kidney disease and thyroid dysfunction, respectively (4). However, in the present study, close to 50% of the data set represented individuals having one or more disaccharidase deficiencies. Since the overall prevalence of disaccharidase deficiency in the population is thought to be low, this data set was skewed toward deficient individuals and not representative of the general population.
Another limitation of this study is the lack of clinical data accompanying the laboratory test results. While the data set appears skewed toward disaccharidase deficiency with respect to the expected prevalence in the population, no information on the ethnic demographics of this data set were available. For instance, the frequency of lactose intolerance, presumably caused by lactase deficiency, is known to vary throughout different populations and different regions of the world (12). Additionally, to truly validate the clinical utility of the proposed modifications to the activity cutoffs to classify deficiency, correlation with clinical symptoms is needed to determine the activity at which a clinical presentation consistent with enzymatic deficiency is present. A study by Gupta et al. (13) reported on such an approach in 232 samples obtained from patients <18 years of age with and without diarrhea and evaluated for mucosal histology and disaccharidase activities. The geometric means of enzyme activities decreased with increasing severity of mucosal injury but were significantly lower only in individuals with diarrhea and moderate to severe histologic changes. Access to biopsy histology and patient symptoms in this study's data set would have been useful and might have altered the outcome of the Hoffman analysis.
The authors thank David Davis and all members of the Special Chemistry Laboratory at ARUP Laboratories.
↵3 Nonstandard abbreviations:
- laboratory information system.
Some parts of this study were previously presented as a poster at the 66th AACC Annual Scientific Meeting, Chicago, IL, July 27–31, 2014.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form.
Employment or Leadership: D.G. Grenache, AACC.
Consultant or Advisory Role: None declared.
Stock Ownership: None declared.
Honoraria: None declared.
Research Funding: None declared.
Expert Testimony: None declared.
Patents: None declared.
Role of Sponsor: No sponsor was declared.
- Received June 27, 2016.
- Accepted June 28, 2016.
- © 2016 American Association for Clinical Chemistry