*Pediatric Gastroenterology Unit and Cystic Fibrosis Center, Hadassah University Hospital, Jerusalem, Israel
†Graub Cystic Fibrosis Center, Schneider Children's Hospital, Petach Tikva, Israel
‡Cystic Fibrosis Center, Israel
||Gastroenterology Laboratory, Shaare Zedek Medical Center, Jerusalem, Israel
§Cystic Fibrosis Center, Rambam Medical Center, Haifa, Israel
¶Cystic Fibrosis Center, Hadassah Medical Center, Jerusalem, Israel
Received 15 October, 2007
Accepted 28 February, 2008
Address correspondence and reprint requests to Michael Wilschanski, MD, Director Pediatric Gastroenterology Unit, Hadassah University Hospital, Mount Scopus, Jerusalem 91240, Israel (e-mail: email@example.com).
The authors report no conflicts of interest.
Objectives: Most patients with cystic fibrosis (CF) have pancreatic insufficiency; however, 15% of the patients are pancreatic sufficient (PS). Several laboratory tests have been developed to distinguish between pancreatic insufficiency and PS. The gold standard to determine pancreatic function apart from direct pancreatic stimulation test is the 72-hour fecal fat excretion, expressed as coefficient of fat absorption (CFA). The aim was to test the correlation between 2 other tests, fecal elastase-1 and serum immunoreactive trypsinogen (IRT), as compared with fecal fat excretion.
Patients and Methods: 21 patients with CF-PS performed the 3 tests of fecal fat excretion, fecal elastase-1, and IRT. Correlation between the tests was evaluated by the κ statistics test, sensitivity and specificity, and positive and negative predictive values.
Results: CFA was abnormal in 5 patients, elastase was <200 μg/g in 4 patients, and IRT was <20 ng/mL in 2 patients. The correlation between CFA and IRT was negative (κ = −0.154), and between CFA and fecal elastase-1 was poor (κ = 0.213). The sensitivity, specificity, and positive and negative predictive values of IRT versus CFA were 0%, 88%, 0%, and 78%, and for fecal elastase-1 were 40%, 81%, 40%, and 81%, respectively.
Conclusions: In CF-PS, poor correlation was found between IRT, fecal elastase-1, and CFA, therefore neither fecal elastase-1 in the stool nor IRT in the serum reaches the sensitivity or the specificity of the fecal fat excretion. Thus, fecal fat excretion is required in patients with CF for evaluation of pancreatic function.
Most patients with cystic fibrosis (CF) experience insufficient pancreatic enzyme excretion leading to fat malabsorption requiring supplementation of pancreatic enzymes to maintain normal growth. These patients are termed pancreatic insufficient (PI), and if not treated with pancreatic enzymes they develop malabsorption, malnutrition, and growth failure. However, approximately 15% of the patients have sufficient pancreatic function to maintain normal nutrition, do not need pancreatic enzyme supplementation, and are termed pancreatic sufficient (PS). A proportion of PS patients eventually develop PI.
The assessment of exocrine pancreatic function is a mandatory procedure at the time of diagnosis, to determine whether the patient should be given pancreatic enzyme replacement therapy (1). Several tests have been developed to determine exocrine pancreatic function. The most precise method is the secretin-cholecystokinin stimulation test, which includes direct duodenal intubation, stimulation with secretin and cholecystokinin, and collection of pancreatic secretions for pH, bicarbonate, and enzyme analysis. This test is the gold standard, but it has major disadvantages because it is invasive, time-consuming, and expensive. As far as indirect pancreatic function is concerned, the gold standard is the 72-hour fecal fat excretion, expressed as the coefficient of fat absorption (CFA). This test is not tolerated well by patients or laboratory personnel, so other indirect tests have been developed. These include breath tests, fecal determination of pancreatic enzymes such as fecal elastase-1, and measurement of serum enzymes such as immunoreactive trypsinogen (IRT), a test used in neonatal screening for CF. Measurement of fecal pancreatic elastase-1 is a noninvasive and simple indirect method to study exocrine pancreatic function, and thus it is widely used. Previous studies have shown that this appears to be a good measure of pancreatic function (2,3); it is highly specific for the pancreas and does not cross-react with pancreatic preparations of animal origin, making it unnecessary to discontinue enzyme replacement therapy in studied subjects (3).
IRT also is used in the assessment of exocrine pancreatic function. It is in widespread use in neonatal screening for CF because IRT levels are elevated in infancy (4). IRT levels fall after 1 year of age and after the age of 7 are in the subnormal range in PI patients. In PS patients, there is no age-related decline in IRT levels, and routine use in PS patients has not been recommended.
The aim of this study is to test the correlation between fecal elastase-1 and IRT as compared with fecal fat excretion and thereby evaluate the best available noninvasive test to determine exocrine pancreatic function in patients with CF.
PATIENTS AND METHODS
Twenty-one patients with CF-PS consented to complete all 3 tests of fecal fat excretion, fecal elastase, and IRT and served as the cohort for this study. All of the patients were diagnosed with CF according to the Rosenstein criteria (5). Patients 1, 2, and 3 are siblings from a Muslim family. They have severe bronchiectasis with mucoid Pseudomonas aeruginosa on sputum culture. Sweat chloride measurements have been normal and no mutation has been identified (60% of Muslim patients). They underwent an extensive workup for other disorders, including immune deficiency and ciliary disorders. They have clubbing and pansinusitis. Nasal potential difference measurements were abnormal on repeated testing. This test measures the transport of sodium and chloride across the apical surface of the epithelial cells and/or the inferior turbinate. An abnormal test according to the Rosenstein criteria (5) may be used as evidence for cystic fibrosis transmembrane conductance regulator dysfunction when sweat chloride results are borderline or normal. Patients 10 and 17 have pansinusitis and an abnormal nasal potential difference measurement. Patients 11 and 12 have clubbing, severe bronchiectasis, and pansinusitis with a family history of CF.
Patients and/or parents were instructed to collect the stool during a period of 72 hours. The diet regimens were standardized for age and weight before and during the 3-day stool collection. A dietitian supervised the diet containing 100 g of fat daily to each adult patient and 50 g to each child for the 2 days before and for the duration of the stool collection. Fecal fat was analyzed by the Van de Kamer method (6). An abnormal result was defined as fecal fat excretion greater than 7% of intake, in other words, less than 93% CFA.
Pancreatic fecal elastase-1 was determined immunologically with an enzyme linked immunosorbent assay (Schebo-Tech, Wettenberg, Germany). The results were expressed as μg fecal elastase-1 per gram of stool. Values below 200 μg/g stool were considered to be abnormal, according to Walkowiak et al (7). Two values of fecal elastase were chosen as thresholds for the calculation of sensitivity and specificity, 100 and 200 μg/g (8). Blood from each patient was taken for serum IRT (DiaSorin, Stillwater, MN). (9).
Sweat chloride measurements, using pilocarpine electrophoresis, and pulmonary function measurements were performed at the referring CF center. All patients had normal vitamin E and A levels. All other laboratory tests were performed at the Gastroenterology Laboratory of Shaare Zedek Medical Center in Jerusalem.
Sensitivity, specificity, and positive and negative predictive values comparing elastase and IRT to CFA were examined. The agreement between the tests was examined by the κ index (10), which compares the agreement against that which may be expected by chance. Kappa is the chance-corrected proportional agreement, and possible values range from 1 (perfect agreement) via 0 (no agreement above that expected by chance) to −1 (complete disagreement). Values between 0.41 and 0.60 are regarded as moderate agreement. Ethics approval was obtained from the Ethics Committee of the Israeli Health Ministry.
Twenty-one patients (median age 11.1 years; range = 9 months–40 years) completed the protocol. Clinical parameters including age, sex, cystic fibrosis transmembrane conductance regulator mutations, and forced expiratory volume in 1 second are presented in Table 1.
The results of the indirect pancreatic function tests are shown in Table 2. Sixteen patients had normal fecal fat excretion and 5 patients had abnormal fecal fat excretion, but all of the children thrived and the adults maintained their body weight without requiring pancreatic enzyme supplementation. Fecal elastase concentrations were 365 ± 190, range 80 to 688; IRT 41 ± 15, range 16 to 78; and CFA 95 ± 3.5, range 88 to 99.
Table 3 shows the statistical analysis of each test individually and in combination. Using a cutoff of 200 μg/g, the sensitivity of fecal elastase-1 was 40%, the specificity was 81%, positive predictive value 40%, and negative predictive value 81%. Using a cutoff of 100 μg/g, the sensitivity of fecal elastase-1 was 94%, the specificity was 20%, positive predictive value 50%, and negative predictive value 79%. The sensitivity of IRT to detect PI was 0%, the specificity was 88%, positive predictive value 0, and negative predictive value 78%.
The κ test showed no correlation between individual tests, so we progressed to using combinations of tests. The term “combined worse” was used if 1 test was positive then the patient was regarded for the analysis only as PI; and “combined best” was defined if 1 test was negative then the patient was regarded as PS. Even with this strategy, the results were not improved.
We have shown that the performance of fecal elastase-1 either alone or even in combination with IRT is not as accurate as the performance of a 3-day fecal fat collection in 21 CF-PS patients. This is the largest number of PS patients reported who have performed all 3 tests. In this population, in which some PS patients may deteriorate to PI, only CFA may be relied on.
Some authors (11) claim that the 72-hour collection of dietary intake and stool for determination of fecal fat content is inconvenient and prone to collection errors in the nonresearch setting. The possibility of error in the reporting of the dietary intake at home and the difficulty to obtain the complete collection of stool restricts the steatorrhea diagnostic value. That study refers primarily to PI and nutritional status, which are not relevant here. However, the performance of a 3-day fecal fat with prior dietary instruction provides a more complete picture of fat balance than a single stool or blood test.
As far as IRT is concerned, most of the studies do not measure sensitivity and specificity. Cleghorn et al (12) found a sensitivity of 93% and specificity of 92% of IRT to detect steatorrhea in 77 CF patients both PS and PI.
Previous studies have shown the relative advantage of fecal elastase-1 as compared with chymotrypsin (7) and lipase (13) in the stool. These studies add further information to previous studies of this group (14), in which fecal elastase was compared with direct pancreatic function testing with cholecystokinin. They claimed that fecal elastase has high sensitivity and specificity but as far as PS patients were concerned, a sensitivity of only 25% was found.
Siegmund et al (15) discussed some of the noninvasive (tubeless) pancreatic function tests used to diagnose exocrine pancreatic insufficiency (PI). Studies evaluating the diagnostic validity of these tests were integrated into a meta-analysis. They found the sensitivity (n = 307, where n = sum of people included in all analyzed studies) of fecal elastase-1 to be 54% for diagnosis of “slight” PI, 75% for “moderate” PI, and 95% for “severe” PI. The specificity was found to be 79% (n = 347). Fecal fat was not included in the analysis. Their conclusion was that none of the noninvasive pancreatic function tests is sensitive enough to diagnose reliably a slight to moderate exocrine pancreatic insufficiency.
Our results regarding fecal elastase-1 are similar, but we do not agree with their conclusions concerning fecal fat collection. The tendency to rely on fecal elastase-1 to differentiate between PS and PI is somewhat troublesome because in previous studies (16) there is up to 37% variability in repeated tests on individual patients.
There is uncertainty as to the normal value of fecal elastase-1 whether it is above 100 or 200 μg/g of stool (8). We performed statistical analysis on both values, and there was no significant change in the result. Our results show that the compatibility (κ) between IRT, fecal elastase-1, and CFA is low. In comparison to CFA, the performances of IRT and fecal elastase-1 do not improve pancreatic function testing. However, the specificity and negative predicted value are high, so it is possible to rule out PI with 80% certainty.
Patient 6 is a good example of this issue. During the study period, she was regarded as PS, because of her normal fecal elastase-1 and IRT, despite her borderline CFA. Some months after the study was finished, she became clinically PI and required enzyme supplements; it may have been to her advantage if she had started them earlier.
There are several possible reasons for these findings. First, all of these tests are indirect pancreatic function tests, hence their variability. We do not claim that any of these tests replace the direct pancreatic stimulation test, which is still performed in some specialist centers (17). Second, these tests are single-point tests, which may be variable from patient to patient, whereas CFA is a longer test and therefore represents a truer picture of pancreatic exocrine function. IRT and fecal elastaste-1 should be performed as a routine screening test in PS patients. We conclude that despite the difficulty in the performance of CFA, there is a definite need to perform this test at time of diagnosis or during follow-up in PS patients, especially if there is a change in clinical status. If the patient is clearly PI, one of these indirect screening tests may be useful, but not in the evaluation and follow-up of a suspected PS patient. The limitation of this study is the relatively small number of patients and the results need to be confirmed in a larger study.
We agree with the conclusions of Walkowiak et al (18) that fecal elastase-1 is a helpful screening tool for the longitudinal assessment of PS patients, but at this time there is no better indirect test for pancreatic exocrine function than CFA. Thus, despite the technical difficulties and the inconvenience of collecting stool, there is a need to evaluate precisely the pancreatic function of the patients with CF, at least at the time of diagnosis, and not infrequently during the follow-up period, especially when changes in clinical parameters occur.
1. Kerem E, Conway S, Elborn S, et al
. Consensus committee. Standards of care for patients with cystic fibrosis: a European consensus. J Cyst Fibros 2005; 4:7–26.
2. Löser C, Möllgaard A, Fölsch UR. Faecal elastase 1: a novel, highly sensitive and specific tubeless pancreatic function test. Gut 1996; 39:580–586.
3. Stein J, Jung M, Sziegoleit A, et al
. Immunoreactive elastase 1: clinical evaluation of a new noninvasive test of pancreatic function. Clin Chem 1996; 42:222–226.
4. Bowling FG, Watson AR, Rylatt DB, et al
. Monoclonal antibody-based enzyme immunoassay for trypsinogen in neonatal screening for cystic fibrosis. Lancet 1987; 11:826–827.
5. Rosenstein BJ, Cutting GR. The diagnosis of cystic fibrosis: a consensus statement. J Pediatr 1988; 132:589–595.
6. Van de Kamer JH, Bokkel Huinink H, Weyers HA. Rapid method for the determination of fat in feces. J Biol Chem 1949; 177:347–355.
7. Walkowiak J, Herzig KH, Strzykala K, et al
. Fecal elastase-1 is superior to fecal chymotrypsin in the assessment of pancreatic involvement in cystic fibrosis. Pediatrics 2002; 110:e7.
8. Beharry S, Ellis L, Corey M, et al
. How useful is fecal pancreatic elastase 1 as a marker of exocrine pancreatic disease? J Pediatr 2002; 141:84–90.
9. Geokas MC, Largman C, Brodrick JW, et al
. Determination of human pancreatic cationic trypsinogen in serum by radioimmunoassay. Am J Physiol 1979; 236:E77–E83.
10. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 8476:307–310.
11. Cohen JR, Schall JI, Ittenbach RF, et al
. Fecal elastase: pancreatic status verification and influence on nutritional status in children with cystic fibrosis. J Pediatr Gastroenterol Nutr 2005; 40:438–444.
12. Cleghorn G, Benjamin L, Corey M, et al
. Serum immunoreactive pancreatic lipase and cationic trypsinogen for the assessment of exocrine pancreatic function in older patients with cystic fibrosis. Pediatrics 1986; 77:301–306.
13. Walkowiak J, Lisowska A, Przyslawski J. Faecal elastase-1 test is superior to faecal lipase test in the assessment of exocrine pancreatic function in cystic fibrosis. Acta Paediatr 2004; 93:1042–1045.
14. Walkowiak J, Cichy WK, Herzig KH. Comparison of fecal elastase-1 determination with the secretin-cholecystokinin test in patients with cystic fibrosis. Scand J Gastroenterol 1999; 34:202–207.
15. Siegmund E, Lohr JM, Schuff-Werner P. The diagnostic validity of non-invasive pancreatic function tests—a meta-analysis. Z Gastroenterol 2004; 42:1117–1128.
16. Meyts I, Wuyts W, Proesmans M. Variability of fecal pancreatic elastase measurements in cystic fibrosis patients. J Cyst Fibros 2002; 1:265–268.
17. Schibli S, Corey M, Gaskin KJ, et al
. Towards the ideal quantitative pancreatic function test: analysis of test variables that influence validity. Clin Gastroenterol Hepatol 2006; 4:90–97.
18. Walkowiak J, Nousia-Arvanitakis S, Agguridaki C, et al
. Longitudinal follow-up of exocrine pancreatic function in pancreatic sufficient cystic fibrosis patients using the fecal elastase-1 test. J Pediatr Gastroenterol Nutr 2003; 36:474–478.
© 2009 Lippincott Williams & Wilkins, Inc.