Baker, Robert D.
Digestive Diseases and Nutrition Center, Women and Children's Hospital of Buffalo, Buffalo, NY.
Address correspondence and reprint requests to Robert D. Baker, MD, PhD, Professor of Pediatrics, University at Buffalo, Digestive Diseases and Nutrition Center, Women and Children's Hospital of Buffalo, 219 Bryant St, Buffalo, NY 14222 (e-mail: rbaker@UPA.CHOB.edu).
Received 26 October, 2012
Accepted 14 November, 2012
The author reports no conflicts of interest.
See “Comparison of Diagnostic Methods for Pancreatic Insufficiency in Infants With Cystic Fibrosis” by Tardelli et al on page 178.
In this issue of JPGN, Tardelli et al (1) report on diagnosing pancreatic insufficiency (PI) in Brazilian infants known to have cystic fibrosis (CF). Establishing the diagnosis of PI is often not easy because of the difficulties in performing and interpreting the tests. The acknowledged criterion standard test for pancreatic function is the measurement of pancreatic enzymes in duodenal fluid before and after intravenous administration of pancreozymin-cholecystokinin-secretin, the pancreatic stimulation test. This test has been used for decades and normal values for pancreatic enzymes were established in the 1950s and 1960s (2,3). The most often used version of this test is the one described by Madrazo-de la Garza et al in 1991 (4). This modified version employs only secretin as a stimulant and uses an endoscope for fluid collection, rather than a triple lumen duodenal tube. The pancreatic stimulation test is the only assessment of pancreatic function that directly measures pancreatic enzymes. Because it is difficult and time-consuming to perform, it is seldom done.
More recently a measure of pancreatic enzyme in stool has been developed and trialed, the fecal elastase test (5). This test has the advantage that it is measured in a single, random stool sample. The test is easy to perform, does not require sophisticated technology, is noninvasive, and is both sensitive and specific. Because the antibody is specific for human elastase, it can be used with concomitant pancreatic enzyme replacement therapy. A major drawback is that it cannot be performed on diarrheal stool, which is often present when one is looking for PI. The accuracy of the test has been questioned in preterm infants and in the first 2 weeks of life in term infants (6).
Other measures of exocrine pancreatic function commonly rely on measuring fat malabsorption. Obviously, any test that measures fat malabsorption will not distinguish between PI and other causes of fat malabsorption, such as primary liver disease or diseases of the gastrointestinal mucosa. Because fat is unevenly distributed in the stool of anyone taking in a mixed diet, spot fecal fat determinations are inaccurate, yielding both false-positive and false-negative results. A modification of the spot fecal fat determination is the steatocrit. The principle of this test is that if a small amount of homogenized stools is centrifuged at 15,000 rpm for 15 minutes in a hematocrit tube, 3 phases will result: lipid phase on top, liquid aqueous phase below, and stool residual phase on the bottom. Sensitivity and specificity of this test are greater in infants receiving a uniform diet as opposed to older individuals receiving a mixed diet and who have uneven distribution of fecal fat. Calculating the coefficient of fat absorption using a measure of fecal fat in a 72-hour stool collection done while maintaining a record of dietary fat intake is the standard test for determining fat malabsorption; however, this test is notoriously difficult for parents and patients to perform and uniformly odious for patients, parents, and laboratory technicians. Therefore, it is seldom done and, when done, is often not done correctly.
There are several indirect measures of pancreatic function that do not depend on fat malabsorption rather on other functions of exocrine pancreatic enzymes. The modified Schilling test relies on the fact that pancreatic enzymes are required to cleave the R factor from cyanocobalamin, and this cleavage is necessary for intrinsic factor-cyanocobalamin binding and subsequent vitamin absorption to occur. The bentiramide test is performed by administering an oral dose of N-benzoyl-L-tyrosyl-p-aminobenzoic acid (bentiramide, Chymex, Savage Labs, Melville, NY). Bentiramide is nonabsorbable until it is cleaved by pancreatic chymotrypsin, yielding N-benzoyl-L-tyrosine and p-aminobenzoic acid. p-Aminobenzoic acid is rapidly absorbed and, after conjugation, is excreted and measured in the urine. The modified Schilling test is rarely done, and the bentiramide test was withdrawn from the US market in 1996. The bentiramide test is still available in some countries.
Although we continue to categorize exocrine pancreatic function as either PI or pancreatic sufficiency (arbitrarily based on whether or not the patient needs pancreatic enzyme replacement therapy to control symptoms), it is now clear that pancreatic function varies along a spectrum from normal to severe deficiency. Over 96% of infants and children with PI have CF; however, not all children with CF have PI. Approximately, 10% of individuals with CF are pancreatic sufficient and the proportion with pancreatic sufficiency decreases with age (7). This means that if all babies with CF are placed on pancreatic enzymes, >1 child in 10 will be receiving medication that he or she does not need. Additionally, knowledge of the pancreatic status of a child with CF allows for prognostication because CF patients with PI have worse nutritional status and more severe lung disease. Thus, it is important to establish the pancreatic status of infants who have CF. Because an infant with CF may be pancreatic sufficient, but may later become PI, repeated tests for pancreatic function may be needed.
The authors of the study appearing in this issue of JPGN compare weight gain plus a 72-hour fecal fat collection or weight gain plus a fecal steatocrit to their chosen criterion standard, fecal elastase. They found weight gain plus a fecal steatocrit was the better gauge of PI and had 91.3% sensitivity and specificity of 83.3% in their young patients. The motivation for the present study appears to be the inability to routinely obtain results of fecal elastase test in the Brazilian setting. Although the authors do not address why fecal elastase is not routine in Brazil, one would suspect that logistics, availability, and costs are the major factors. As alluded to in this report, steatocrit combined with weight gain is reliable only in infants. Beyond infancy, reasons for poor weight gain become overwhelmingly numerous and because of a mixed diet steatocrit becomes less accurate; therefore, for follow-up testing, even in the Brazilian setting, a test other than steatocrit will be necessary.
The optimal treatment for all infants with CF is to prevent malnutrition from occurring. This goal requires knowing pancreatic function status. In the long run, determining fecal elastase will best accomplish this. My advice for Brazil and other areas where CF is encountered is to establish a mechanism for implementing routine fecal elastase testing. For the present, where fecal elastase determination is not feasible, weight gain plus steatocrit appears to be a reasonable alternative for infants.
1. Tardelli ACS, Camargos PAM, Penna FJ, et al. Comparison of diagnostic methods for pancreatic insufficiency in infants with cystic fibrosis. J Pediatr Gastroenterol Nutr 2013;56:178–81.
2. Hummel BCW. A modified spectrophotometric determination of chymotrypsin, trypsin and thrombin. Can J Biochem Physiol 1959; 37:1393–1399.
3. Dahlquist A. A method for the determination of amylase in intestinal content. Scand J Clin Lab Invest 1962; 14:145–151.
4. Madrazo-de la Garza JA, Gotthold M, Lu B, et al. A new direct pancreatic function test in pediatrics. J Pediatr Gastroenterol Nutr 1991; 12:356–360.
5. Borowitz D, Baker SS, Duffy L, et al. Use of fecal elastase-1 to classify pancreatic status in patients with cystic fibrosis. J Pediatr 2004; 145:322.
6. Nissler K, Von Katte I, Huebner A, et al. Pancreatic elastase 1 in feces of preterm and term infants. J Pediatr Gastroenterol Nutr 2001; 33:28–31.
7. Walkowiak J, Sands D, Nowakowska A, et al. Early decline of pancreatic function in cystic fibrosis patients with class 1 or 2 CFTR mutations. J Pediatr Gastroenterol Nutr 2005;40:199–201.