Insulin and Glucose Excursion Following Premeal Insulin Lispro Or Repaglinide in Cystic Fibrosis–Related Diabetes

Clinical Care/Education/Nutrition ORIGINAL ARTICLE

Insulin and Glucose Excursion Following Premeal Insulin Lispro or Repaglinide in Cystic Fibrosis–Related Diabetes

1 2 ANTOINETTE MORAN, MD CARLOS MILLA, MD Patients with CF-related diabetes (CFRD) 2 JAMES PHILLIPS, MD are more underweight and have worse pulmonary function than those without diabetes (5). Clinical decline may be di- rectly related to the diabetic state and to insulin deficiency. Insulin therapy has OBJECTIVE — Insulin and glucose levels in response to premeal insulin lispro or repaglinide been found to reverse negative changes in were evaluated in adult patients with cystic fibrosis–related diabetes (CFRD) without fasting weight and lung function in CFRD, sug- hyperglycemia. gesting a cause-and-effect relationship between insulin deficiency and CF clini- RESEARCH DESIGN AND METHODS — Seven patients with CFRD were fed 1,000- kcal liquid mixed meals. Three study conditions were administered in random order on separate cal deterioration (6,7). More substantial mornings: 1) no premeal diabetes medication, 2) insulin lispro, 0.1 unit/kg body wt premeal and evidence for this relationship is provided 3) repaglinide 1 mg premeal. Glucose and insulin levels were measured every 20 min for 5 h. by the observation that both the severity of glucose intolerance and the degree of RESULTS — Fasting insulin and glucose levels were normal in patients with CFRD, but the insulin deficiency in an individual with peak glucose level was elevated. Insulin lispro significantly decreased the peak glucose level (P ϭ CF correlate with future decline in pul- 0.0004) and the 2-h (P ϭ 0.001) and 5-h (P Ͻ 0.0001) glucose area under the curve (AUC). monary function (8). Weight loss and ϭ Repaglinide significantly decreased the 5-h glucose AUC (P 0.03). Neither drug completely more severe pulmonary disease in CF are normalized cystic fibrosis glucose excursion at the doses used for this study. Insulin lispro ϭ associated with protein catabolism significantly increased the 5-h insulin AUC (P 0.04). (9,10), and protein catabolism may be re- CONCLUSIONS — In response to subcutaneous insulin lispro, postprandial glucose excur- lated to the degree of glucose intolerance sion was significantly diminished and insulin secretion was enhanced compared with a control (11). We have hypothesized that insulin meal in which no medication was given to patients with CFRD. The oral agent repaglinide deficiency in CF leads to protein catabo- resulted in lesser corrections in these parameters. Neither drug completely normalized glucose lism, thereby negatively influencing or insulin levels, suggesting that the doses chosen for this study were suboptimal. Placebo- weight, pulmonary function, and ulti- controlled longitudinal studies comparing the effectiveness of repaglinide and insulin on glucose mately, survival (12). Therefore, it is im- metabolic control as well as overall nutrition and body weight are needed to help determine portant to determine how to most optimal medical treatment of CFRD. effectively treat CFRD. Current recommendations for the Diabetes Care 24:1706–1710, 2001 management of CFRD are based on clinical experience rather than experimental evidence because so few data exist in the iabetes and impaired glucose toler- tolerance, and 25% normal glucose toler- literature. In particular, there is consider- ance are common complications of ance (1). Diabetes is primarily due to in- able controversy regarding the treatment D cystic fibrosis (CF). Based on rou- sulin deficiency (2), which is caused by of CFRD without fasting hyperglycemia tine annual oral glucose tolerance test fibrotic destruction of pancreatic islets (13). In the current study, we evaluated screening of the 450 patients with CF fol- and perhaps by dysfunction of the re- the acute plasma insulin and glucose re- lowed at the University of Minnesota, we maining ␤-cells related to islet amyloid sponses to premeal injection of insulin reported a prevalence in adults with CF deposition (3). lispro or to oral administration of repa- of ϳ15% diabetes with fasting hyper- In CF, the additional diagnosis of di- glinide in adult patients with CF and dia- glycemia, 25% diabetes without fasting abetes is associated with significantly in- betes without fasting hyperglycemia. hyperglycemia, 35% impaired glucose creased morbidity and mortality (4–6). ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● RESEARCH DESIGN AND 1 2 From the Divisions of Endocrinology and Pulmonology, Department of Pediatrics, University of Minne- METHODS sota, Minneapolis, Minnesota. Address correspondence and reprint requests to Antoinette Moran, MD, Department of Pediatrics, MMC 404, University of Minnesota, 516 Delaware St., Minneapolis, MN 55455. E-mail: [email protected]. Subjects Received for publication 6 April 2001 and accepted in revised form 22 June 2001. Seven patients with CF and a routine oral Antoinette Moran has received a research grant from Novo Nordisk Pharmaceuticals to test repaglinide glucose tolerance test in the previous 3 and Novolog insulin in cystic fibrosis. Abbreviations: AUC, area under the curve; CF, cystic fibrosis; CFRD, cystic fibrosis–related diabetes. months who demonstrated CFRD with- A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion out fasting hyperglycemia were recruited factors for many substances. into the study. Patients were excluded

1706 DIABETES CARE, VOLUME 24, NUMBER 10, OCTOBER 2001 Moran, Phillips, and Milla

Table 1—Characteristics of patients with CFRD without fasting hyperglycemia mean Ϯ SEM. Data were analyzed for different responses to the study conditions ␣ CFRD by two-way analysis of variance. An value of 0.05 was used as the cutoff for statisti- n 7 cal significance. The SAS statistical soft- Age, years 24 Ϯ 5 ware package (SAS Institute, Cary, NC) Sex, F/M 4/3 was used for the statistical analysis. Based Height, inches 64 Ϯ 3 on data from previous studies (1), we es- Weight, kg 55 Ϯ 7 timated that with a sample of seven pa- BMI, kg/m2 21 Ϯ 1 tients studied per condition, we would have Fasting plasma glucose, mg/dl (mmol/l) 98 Ϯ 18, 5.4 Ϯ 1.0 a Ͼ80% chance of detecting significant dif- Fasting insulin, ␮U/ml (pmol/l) 5 Ϯ 1, 30 Ϯ 6 ferences in the glucose excursion values at Ϯ HbA1c, % 5.6 0.9 (range 3.9–6.6) the chosen ␣ value for significance. Data are n, means Ϯ SD, and means Ϯ SD (range). RESULTS who had a history of CF liver disease or (unpublished data). Patients with CF took who had experienced acute illness or un- their usual dose of digestive enzymes be- Subjects dergone oral or intravenous glucocorti- fore consuming this meal; all other morn- Characteristics of the subjects with CFRD coid therapy during the previous 3 ing medications and treatments were held are presented in Table 1. None of the pa- months. Because the study subjects with until after the study day was complete. On tients with CF had signs or symptoms of CF were chronically infected with CF each day, one of three study conditions acute illness in the 3 months preceding pathogens, they all received chronic sup- was administered in random order: 1)no the study. All of them were engaged in pressive aerosolized antibiotic prepara- premeal diabetes medication (baseline relatively normal life styles and were ei- tions. Seven healthy, nonathlete, age-, meal); 2) insulin lispro, 0.1 unit/kg body ther college students or employed at least sex- and BMI-matched normal control wt, subcutaneous injection in the poste- part-time. Controls were matched for age, subjects were recruited by poster adver- rior upper arm using a short (3/8-inch) sex, and BMI. tisement. Approval for this study was ob- needle syringe, 10 min premeal; and 3) tained from the University of Minnesota repaglinide 1 mg orally, 10 min premeal. Response to the baseline meal Committee for the Use of Human Subjects Typical clinical practice for initiating without medication in Research; informed consent was ob- insulin therapy in clinically stable pa- Fasting insulin and glucose levels were tained from all subjects. tients with CFRD involves giving a pre- normal in patients with CFRD (Table 1). meal insulin dose of 0.1 unit insulin per Postprandial glucose excursion was con- kg body wt or giving 0.5 unit insulin per Study design siderably elevated in the patients with CF, 15 g carbohydrate (14). The current pro- with a peak glucose level of 228 Ϯ 30 Subjects with CF were studied at the Uni- tocol, based on weight, resulted in insulin Ϯ versity of Minnesota General Clinical Re- mg/dl (12.7 1.7 mmol/l) (Fig. 1). As doses of 5–7 units. If the dose had been cal- previously reported for patients with CF search Center on three separate occasions culated based on carbohydrate counting, during a 1- to 2-month period. They were (2), the peak insulin levels were markedly each patient would have received 4 units. delayed and blunted (Fig. 1). admitted in the evening and not allowed Control subjects received the 1,000- to eat or drink after 10:00 P.M. An indwell- kcal test meal under the same conditions Response to insulin lispro ing catheter was inserted retrogradely into on a single occasion with no premeal In response to premeal insulin lispro, pa- a dorsal hand vein for blood sampling, medication. Plasma insulin and glucose tients with CF experienced a reduction in and the hand was placed on a heating pad levels were measured at 20-min intervals glucose excursion (Table 2). Compared during the studies. for 5 h. For reporting purposes, time 0 is with the untreated baseline meal, im- On each occasion, subjects were fed a considered the beginning of the meal (10 A M provements were seen in the peak glucose 1,000-kcal liquid mixed meal at 7:00 . . min after medications were given). The meal was consumed within 30 min. level (P ϭ 0.0004), the 2-h glucose area The mixed meal, prepared by the Uni- under the curve (AUC) (P ϭ 0.001), and versity of Minnesota General Clinical Analytical methods the 5-h glucose AUC (P Ͻ 0.0001). De- Research Center dietary staff, had a milk- Plasma glucose was measured immedi- spite these improvements, glucose excur- shake-like consistency and a macronutri- ately with a Beckman Glucose Analyzer II sion was still abnormally elevated in CF ent composition of 47% carbohydrate, (Beckman Instruments, Fullerton, CA). (Fig. 1). 17% protein, and 36% fat. Although the Insulin samples were collected on ice and Premeal insulin lispro significantly percentage of calories derived from fat centrifuged within 20 min. Insulin levels enhanced the insulin AUC for the 5-h was somewhat less than that recom- were determined by radioimmunoassay study period (65% increase, P ϭ 0.04) mended for individuals with CF (40%), using a double-antibody method. (Table 2). Although there was a trend for the calorie content and macronutrient an increase in the insulin AUC during the composition are typical of meals recorded Statistical analysis first 2 h, this did not achieve statistical by patients with CF at the University of Baseline data are reported as mean Ϯ SD significance (64% increase, P ϭ 0.10). Minnesota during routine diet histories (Table 1); all other data are presented as The peak insulin level was not signifi-

DIABETES CARE, VOLUME 24, NUMBER 10, OCTOBER 2001 1707 Insulin lispro and repaglinide in CFRD

the two drugs in the peak glucose level (P ϭ 0.02), the 2-h glucose AUC (P ϭ 0.02), and the 5-h glucose AUC (P ϭ 0.01). However, we cannot exclude the possibility that higher doses of repaglinide might have resulted in greater improvements. Insulin lispro signiﬁcantly increased the 5-h insulin AUC in CF (P ϭ 0.04). Neither drug, at the doses used in the present study, signiﬁcantly changed the peak insulin level or the 2-h insulin AUC compared with baseline.

Adverse events Four episodes of hypoglycemia (glucose level 48–54 mg/dl, 2.7–3.0 mmol/l) occurred in patients with CF during this Figure 1—Mean plasma insulin (A) and glucose (B) levels in patients with CFRD in response to study: one after the test meal without the baseline test meal with no medication and in response to either subcutaneous 0.1 unit/kg insulin medication, two after administration of lispro or 1 mg oral repaglinide, given 10 min before the mixed meal began at time 0. Control insulin lispro, and one after administra- patients were studied in response to a baseline meal only, with no medication. tion of repaglinide. In all cases, cate- cholamine-related symptoms were cantly different from that seen during the ide, insulin levels seemed to rise (Fig. 1), present but were easily tolerated, and pa- baseline meal (P ϭ 0.37). although there was no significant change tients were alert, oriented, and aware that compared with the baseline meal in the they were hypoglycemic. Hypoglycemia ϳ Response to repaglinide peak insulin level, the 2-h insulin AUC, or occurred an average of 4 h after the test In response to premeal oral repaglinide, the 5-h insulin AUC (all P Ͼ 0.20). meal and resolved spontaneously without glucose excursion was somewhat improved glucose administration after 10–15 min. in CFRD (Table 2). The 5-h glucose AUC Comparison of insulin lispro and was significantly less than baseline (P ϭ repaglinide in CFRD CONCLUSIONS — In patients with 0.03). No significant differences were Insulin lispro seemed to have a more ben- CFRD who are not acutely ill, basal insu- seen in the 2-h glucose AUC (P ϭ 0.39) or eficial effect than repaglinide on post- lin needs are fairly low and hyperglycemia the peak glucose level (P ϭ 0.17). prandial glucose excursion in CFRD. is predominantly associated with meal After administration of 1 mg repaglin- Significant differences were seen between carbohydrate consumption (14). Premeal

Table 2—Plasma glucose and insulin responses in CFRD to a baseline 1,000-kcal liquid test meal and when either 0.1 unit/kg body wt insulin lispro or 1 mg repaglinide were given 10 min before the meal

Baseline meal (no medication) Insulin lispro (0.1 unit/kg) Repaglinide (1 mg) Glucose, mg/dl (mmol/l) Time to peak level (min) 74 Ϯ 766Ϯ 860Ϯ 7 Peak level 228 Ϯ 30 172 Ϯ 9*† 208 Ϯ 18 (12.7 Ϯ 1.7) (9.6 Ϯ 0.5) (11.6 Ϯ 1.0) 2-h AUC 19,574 Ϯ 1,162 14,980 Ϯ 1,056*† 18,247 Ϯ 1,559 (1,087 Ϯ 65) (832 Ϯ 59) (1,014 Ϯ 87) 5-h AUC 40,190 Ϯ 2,121 29,859 Ϯ 1,653*† 35,212 Ϯ 2,067* (2,233 Ϯ 118) (1,658 Ϯ 92) (1,956 Ϯ 115) Insulin, ␮U/ml (pmol/l) Time to peak level (min) 117 Ϯ 11 94 Ϯ 17 107 Ϯ 13 Peak level 53 Ϯ 11 75 Ϯ 970Ϯ 11 (318 Ϯ 66) (450 Ϯ 64) (420 Ϯ 66) 2-h AUC 3,050 Ϯ 709 5,006 Ϯ 676 4,298 Ϯ 716 (18,300 Ϯ 4,254) (30,036 Ϯ 4,056) (25,788 Ϯ 4,296) 5-h AUC 6,446 Ϯ 1,082 10,643 Ϯ 1,753* 8,948 Ϯ 1,162 (38,676 Ϯ 6,492) (63,858 Ϯ 10,518) (53,688 Ϯ 6,972) Data are expressed as means Ϯ SEM. *P Ͻ 0.05 compared with baseline meal; †P Ͻ 0.05 insulin versus repaglinide. Time 0 is considered the beginning of the meal, 10 min after administration of medication.

1708 DIABETES CARE, VOLUME 24, NUMBER 10, OCTOBER 2001 Moran, Phillips, and Milla insulin, given to control postprandial glu- available, these drugs should only be con- meal or nocturnal hypoglycemia. Theo- cose excursion, has become the corner- sidered within the context of controlled retically, this profile of action is ideal for stone of diabetes management in this research trials (13). The literature consists the patient with CFRD without fasting hy- population. Until a few years ago, regular primarily of abstracts and uncontrolled perglycemia. insulin was the only short-acting insulin studies with small subject numbers (18– Clearly, the quality of life and the available for premeal therapy. Anecdot- 21). There are potential problems with complexity of the medical regimen for pa- ally, we found that it frequently caused most of the existing oral diabetes agents tients with CFRD would be improved if hypoglycemia several hours after admin- approved by the Food and Drug Admin- multiple daily injections could be re- istration in patients with CF. This was not istration. Because the primary problem in placed by an oral agent. Our data on the surprising, because it is known that insu- CF is insulin deficiency, drugs that im- acute effects of repaglinide suggest that lin levels may be elevated for 4–6 h after prove insulin sensitivity are not likely to this drug may have therapeutic potential regular insulin injection, even if blood be of major benefit. Metformin is contra- in CFRD. Improvements in glucose ex- glucose has returned to a fasting level. In- indicated in patients with hypoxia be- cursion and plasma insulin levels were sulin lispro was approved for clinical use cause of the risk of fatal lactic acidosis, modest compared with the response to in the U.S. in 1996. It has more rapid and it has multiple gastrointestinal side insulin lispro. However, higher doses of absorption, faster onset, and shorter du- effects that may negatively influence the repaglinide might have resulted in a ration of action than regular human insu- nutritional status of patients with CF. The greater effect. The manufacturer recom- lin, although its potency is equivalent to potential for serious hepatic toxicity with mends a starting dose of 0.5 mg for pa- Ͻ that of regular insulin and the systemic thiazolidinediones may be greater in CF tients whose HbA1c is 8.0 (which would clearance rates of both types of insulin are than in the general population because include most patients with CFRD), but similar. In several studies of patients with underlying liver damage is common. doses up to 4 mg can be prescribed. It is type 1 and type 2 diabetes, use of insulin Drugs that reduce postprandial glucose well established that patients with CF, lispro as the premeal insulin led to signif- excursion by limiting intestinal absorp- who have decreased ␤-cell mass and lack icantly less hypoglycemia than use of pre- tion of glucose cannot be recommended a first-phase insulin response, have de- meal regular human insulin. Because of to patients who suffer from chronic mal- layed insulin secretion (2). Because repa- this and based on clinical experience, in- absorption and undernutrition. glinide acts by stimulating endogenous sulin lispro has been recommended for Sulfonylureas have been the most insulin secretion, it is not surprising that premeal insulin coverage in CFRD (14). commonly used oral agents in CF. Gly- the insulin response to this oral agent fol- The present study was performed to buride, however, is not recommended, lowed the same delayed pattern as the objectively characterize the acute glucose because 50% of its elimination is in the natural response to the baseline meal. and insulin responses to insulin lispro bile. In the few studies that measured the Glycemic control is not the only con- and the oral agent repaglinide in the effect of sulfonylureas on insulin secre- cern with diabetes treatment in CF. Main- CFRD population. Glucose and insulin tion in CF, minimal improvement, if any, tenance of normal body weight is critical excursion were significantly improved was found (19–22). There are theoretical for survival in CF. Although the long- with administration of insulin lispro, and concerns that these agents may impair the term effect of repaglinide on blood glu- a greater effect would likely have been effectiveness of new drugs designed to cose metabolic control is an important seen with a larger dose. Insulin absorp- target the basic defect in CF by improving consideration, the question of whether re- tion seemed to be delayed, because peak function of the CF transmembrane con- paglinide treatment is able to match the insulin levels after injection of insulin lis- ductance regulator, because sulfonylureas anabolic effect of insulin therapy on pro are reported to occur at 41–61 min in bind to and inhibit the CF transmembrane weight gain in CFRD may ultimately be individuals with type 1 diabetes (15,16) conductance regulator (23). The main far more significant. and in normal control subjects (17), practical concern with sulfonylureas has In summary, the current study dem- whereas the peak insulin level in patients been hypoglycemia, which has anecdot- onstrates the effectiveness of insulin lis- with CF in the present study occurred at ally been found to be a common side ef- pro in acutely improving postprandial 94 Ϯ 17 min. The factors controlling in- fect in CF, even with minimal drug doses, glucose and insulin levels. The oral agent sulin absorption are complex. There are limiting the ability to achieve therapeutic repaglinide resulted in lesser improve- many abnormalities in CF that might alter doses. ments in these parameters. The doses of insulin absorption kinetics, including in- Of the currently available oral agents, insulin and repaglinide, although based terstitial fluid pH and electrolyte compo- repaglinide and related agents seem to hold on standard practice recommendations, sition, structural changes in CF adipose the greatest promise in CF. Repaglinide seemed to be too low; therefore, higher tissue, vascular changes brought about by lowers blood glucose levels by stimulat- doses of these drugs may have had great- chronic hypoxia, or increased insulin ing insulin secretion, but in contrast to er therapeutic effect. At present, based degradation in the setting of chronic in- sulfonylureas, insulin release is glucose- on clinical experience, rapid-acting in- flammation. Further studies are needed to dependent and diminishes at low glucose sulin remains the standard premeal di- clarify the question of whether insulin ab- concentrations (repaglinide product in- abetes treatment for patients with CFRD. sorption is delayed in CF. sert; Novo Nordisk Pharmaceuticals, Placebo-controlled longitudinal studies The use of oral agents to treat CFRD Princeton, NJ). This, combined with its comparing the effectiveness of repagli- remains controversial, and a national short half-life (1 h), promotes improve- nide and insulin on not only glucose consensus conference on CFRD rec- ment of meal-stimulated insulin secretion metabolic control but also on overall nu- ommended that, until further data are while minimizing the chance of between- trition and body weight are needed to

DIABETES CARE, VOLUME 24, NUMBER 10, OCTOBER 2001 1709 Insulin lispro and repaglinide in CFRD help determine the optimal medical treat- cose intolerance at baseline. Am J Resp Crit daco P, Di Vincenzo A, Rambotti AM, Mo- ment of CFRD. Care Med 162:891–895, 2000 darelli C, Epifano L, Kassi G, Perreillo G, 9. Steinkamp G, von der Hardt H: Improve- Brunetti P, Bolli G: Effects of the short- ment of nutritional status and lung acting insulin analog [Lys(B28),Pro(B29)] Acknowledgments— This work was sup- function after long-term nocturnal gas- on postprandial blood glucose control in ported by grants from the Cystic Fibrosis trostomy feedings in cystic fibrosis. J Pe- IDDM. Diabetes Care 19:945–952, 1996 Foundation, the National Institutes of Health diatr 124:244–249, 1994 17. Howey DC, Bowsher RR, Brunelle RL, (grant M01-RR-00400, General Clinical Re- 10. Shepherd RW, Holt TL, Thomas BJ, Kay Woodworth JR: Lys(B28), Pro(29)-hu- search Center), and the Minnesota Medical L, Isles A, Francis PJ, Ward LC: Nutri- man insulin: a rapidly absorbed analogue Foundation. tional rehabilitation in cystic fibrosis: of human insulin. Diabetes 43:396–402, controlled studies of effects on nutritional 1994 growth retardation, body protein turn- 18. Bertele-Harms RM, Harms HK: Sulfonyl- References over, and course of pulmonary disease. 1. Moran A, Doherty L, Wang X, Thomas W: J Pediatr 109:788–794, 1986 urea in the treatment of CFRD: a 15 year Abnormal glucose metabolism in cystic 11. Hardin DS, Leblanc A, Lukenbaugh S, experience (Abstract). Pediatr Pulmonol fibrosis. J Pediatr 133:10–16, 1998 Para L, Seilheimer DK: Proteolysis associ- 13:380A, 1996 2. Moran A, Diem P, Klein DJ, Levitt MD, ated with insulin resistance in cystic fibro- 19. Culler FL, McKean LP, Buchanan CJ, Robertson RP: Pancreatic endocrine func- sis. Pediatrics 101:433–437, 1998 Caplan DB, Meacham LR: Glipizide treat- tion in cystic fibrosis. J Pediatr 118:715– 12. Moran A, Milla C, DuCret R, Nair KS: Pro- ment of patients with CF and impaired 723, 1991 tein metabolism in clinically stable adult glucose tolerance. In Endocrine Society 3. Couce M, O’Brien TD, Moran A, Roche CF patients with abnormal glucose toler- Program and Abstracts, San Antonio, TX, PC, Butler PC: Diabetes mellitus in CF is ance. Diabetes 50:1336–1343, 2001 1992. San Antonio, TX, Endocrine Soci- characterized by islet amyloidosis. J Clin 13. Moran A, Hardin D, Rodman D, Allen HF, ety, 1992, p. 45A Endocrinol Metab 81:1267–1272, 1996 Beall RJ, Borowitz D, Brunzell C, Camp- 20. Hinds A, Sheehan AG, Parsons HG: To- 4. Finkelstein SM, Wielinski CL, Elliott GR, bell PW, Chesrown SE, Duchow C, Fink butamide causes a modest increase in Warwick WJ, Barbosa J, Wu SC, Klein DJ: RJ, FitzSimmons SC, Hamilton N, Hirsch insulin secretion in CF patients with im- Diabetes mellitus associated with cystic I, Howenstine MS, Klein DJ, Madhun Z, paired glucose tolerance. Metabolism 44: fibrosis. J Pediatr 112:373–377, 1988 Pencharz PB, Quittner AL, Robbins MK, 13–18, 1995 5. Cystic Fibrosis Foundation: Patient Regis- Schindler T, Schissel K, Schwarzenberg 21. Horswill CA, Zipf WB, Kien CL, McCoy try: Annual Data Report. Bethesda, MD, SJ, Stallings VA, Tullis E, Zipf WB: Diag- KS, O’Dorisio TM: Glipizide increased in- Cystic Fibrosis Foundation, 1998 nosis, screening, and management of sulin secretion but not growth of children 6. Lanng S, Thorsteinsson B, Nerup J, Koch CFRD: a consensus conference report. Di- with CF (Abstract). FASEB J 4:3272A, C: Influence of the development of diabe- abetes Res Clin Pract 45:55–71, 1999 1990 tes mellitus on clinical status in patients 14. Moran A: CFRD: an approach to diagnosis with cystic fibrosis. Eur J Pediatr 151:684– and management. Pediatr Diabetes 1:14– 22. Zipf WB, Kien CL, Horswill CA, McCoy 687, 1992 18, 2000 KS, O’Dorisio T, Pinyerd BL: Effects of 7. Lanng S, Thorsteinsson B, Nerup J, Koch 15. Heinemann L, Heise T, Wahl LC, Traut- tolbutamide on growth and body compo- C: Diabetes mellitus in cystic fibrosis: mann ME, Ampudia J, Starke AAR, Berger sition of nondiabetic children with CF. effect of insulin therapy on lung func- M: Prandial glycaemia after a carbohy- Pediatr Res 30:309–314, 1991 tion and infections. Acta Paediatr 83:849– drate-rich meal in type 1 diabetic patients: 23. Sheppard DJ, Welsh MJ: Effect of ATP- 853, 1994 using the rapid acting insulin analogue sensitive Kϩ channel regulators on cystic 8. Milla CE, Warwick WJ, Moran A: Trends [Lys(B28), Pro(B29)] human insulin. Dia- fibrosis transmembrane conductance reg- in pulmonary function in cystic fibrosis bet Med 13:625–629, 1996 ulator chloride currents. J Gen Physiol 100: patients correlate with the degree of glu- 16. Torlone E, Pampanelli S, Lalli C, Del Sin- 573–591, 1992

1710 DIABETES CARE, VOLUME 24, NUMBER 10, OCTOBER 2001