Effect of Local Corticosteroid Injection of the Hand and Wrist on Blood Glucose in Patients With Diabetes Mellitus

Transcription

1 Effect of Local Corticosteroid Injection of the Hand and Wrist on Blood Glucose in Patients With Diabetes Mellitus Louis W. Catalano III, MD; Steven Z. Glickel, MD; O. Alton Barron, MD; Richard Harrison, MD; Astrid Marshall, BA; Marissa Purcelli-Lafer, BA abstract Full article available online at Healio.com/Orthopedics. Search: Locally administered corticosteroids are a common therapy in many hand and wrist disorders. Corticosteroids pose a theoretical risk to patients with diabetes mellitus by potentially raising blood glucose to hyperglycemic levels. Although oral corticosteroids are known to have an effect on blood glucose control, limited data exist on extra-articular administration. The purpose of this study was to examine the systemic impact of extra-articularly administered corticosteroids in the hand and wrist on serum glucose concentration in patients with diabetes mellitus. Twenty-three patients with diabetes mellitus received a 1-mL triamcinolone acetonide injection for de Quervain s tenosynovitis, trigger finger, flexor carpi ulnaris tendonitis, or carpal tunnel syndrome. Patients recorded their daily morning blood glucose levels for 1 week before injection and for 4 weeks after injection. Average blood glucose levels increased slightly from baseline after injection, reaching statistical significance 1, 5, and 6 days after injection, but were not clinically significant (average increase, 14.2, 9.7, and 32.7 mg/dl, respectively). Isolated increases more than 2 times the standard deviation of preinjection values occurred at least once in the majority of patients. The frequency of hyperglycemic episodes increased after injection, but the proportions of patients with at least 1 hyperglycemic episode before and after injection were not significantly different. These results suggest that local corticosteroid injections are a clinically safe treatment option for inflammatory processes of the hand and wrist in patients with diabetes mellitus. On average, patients experienced slight increases in blood glucose after receiving an injection. Most experienced isolated increases substantially beyond baseline and isolated hyperglycemic effects, but these did not pose an apparent clinical risk. Drs Catalano, Glickel, Barron, and Harrison and Mss Marshall and Purcelli-Lafer are from the C.V. Starr Hand Surgery Center, New York, New York. Drs Catalano, Glickel, Barron, and Harrison and Mss Marshall and Purcelli-Lafer have no relevant financial relationships to disclose. Correspondence should be addressed to: Louis W. Catalano III, MD, C.V. Starr Hand Surgery Center, th Ave, New York, NY (cvstarr@gmail.com). doi: / e1754

2 Effect of Corticosteroid Injection on Blood Glucose Catalano et al Locally administered steroids are a common first-line therapy in many upper-extremity disorders, including trigger finger, dequervain s tenosynovitis, medial and lateral epicondylitis, bursitis, and rotator cuff tendonitis. Steroid injections have been shown to be efficacious in treating these problems and at times providing the definitive cure. 1-3 Patients with diabetes mellitus have a higher incidence of musculoskeletal disorders of the hand compared with age- and sex-matched controls. 4,5 These patients have been shown to have an increased incidence of trigger finger, with a 10% lifetime risk compared with 2.5% in healthy controls. 4,6 In addition to having a higher incidence of such conditions, patients with diabetes mellitus have been shown to be less responsive to steroid injections. 1,4,7,8 However, the first line of treatment is often local corticosteroid injection because of the ease of delivery and, presumably, low risk to the patient. However, corticosteroids pose a theoretical risk to patients with diabetes mellitus. Evidence exists that locally administered corticosteroids may have a systemic effect, which, in patient with diabetes mellitus, may result in elevated blood glucose levels. 9 Wang and Hutchinson 10 reported that a local injection of methylprednisolone in trigger fingers created a hyperglycemic effect that persisted for at least 5 days. They concluded that the transient effect was low risk when considering that corticosteroid injection prevented the need for surgery in nearly half of these patients. They recommended local steroid injection as the primary treatment of trigger finger in patients with diabetes mellitus. 10 The purpose of the current study was to further elucidate the systemic effect of extra-articularly administered corticosteroids in the hand and upper extremity on serum glucose concentration in patients with diabetes mellitus. After observing a patient with diabetes mellitus who became hyperglycemic (blood glucose more than 500 mg/dl) after receiving an extra-articular corticosteroid injection, the authors hypothesized that such events may occur with greater-than-expected frequency. Materials and Methods Fifty-four patients presenting with more than a 6-month history of insulin- and noninsulin-dependent diabetes mellitus who elected to receive a corticosteroid injection of the hand or wrist were included in the study s prospective cohort. Patients taking inhaled or oral steroids for other conditions, patients with a diagnosis of gestational diabetes only, and patients without blood glucose values for the week before injection recorded in a personal log were excluded. If not tested in the previous 30 days, hemoglobin A1c (HbA1c) was measured. Patients were injected with a combination of 1 ml of triamcinolone acetonide and 1 ml of 2% lidocaine, with the exception of 1 patient injected with 1 ml of triamcinolone acetonide and 3 ml of 2% lidocaine. Patients recorded their morning, noon, evening, and bedtime blood glucose levels for 1 week before injection and for 4 weeks after injection. Triamcinolone acetonide has been shown to have a serum half-life of 3.2 to 6.4 days and an average duration of action of 14 days. 11,12 Although a single intra-articular injection of 10 mg of triamcinolone acetonide was shown to be detectable in plasma for more than 2 weeks, data on the half-life of triamcinolone acetonide after extra-articular injection into the tendon sheath are limited 11 ; therefore, preferring not to cut the study period too short, the current authors set the postinjection observation period of this study for 4 weeks. All patients monitored their own blood glucose levels by finger stick with home monitoring devices. All patients administered their own medication, whether oral or insulin by injection. Patients did not specifically record whether blood glucose levels were preprandial vs postprandial. Because it is common when monitoring diabetes for patients to check their blood glucose before breakfast, 13 the authors understood the morning blood glucose values provided by patients to be fasting; however, they could not be certain. Institutional review board approval was obtained for this study. Informed consent and Health Insurance Portability and Accountability Act consent were obtained from each patient. Two main statistical analyses were performed. First, patients median preinjection morning blood glucose values were compared with their morning blood glucose values on each day after injection by Wilcoxon signed rank test for paired samples. Likewise, patients median preinjection bedtime blood glucose values were compared with their postinjection bedtime blood glucose values by Wilcoxon signed rank test for paired samples. However, this approach uses aggregate data, comparing averages, and the authors felt the data were too variable to reveal significance. As an alternate strategy to examine significant blood glucose increases for each patient as opposed to the aggregate data, the authors defined a significant postinjection increase as one more than 2 times the standard deviation of preinjection values, a strategy Habib and Miari 14 used to analyze blood glucose for increases after intra-articular injections of the knee. This way, one looks at each value for each patient and determines whether that value is significantly higher than that patient s baseline values. The frequency of such significant increases was documented. Second, hyperglycemic events (blood glucose more than 180 mg/dl, per clinical opinion of the endocrinology department) were considered to be clinically significant. The number of patients reporting a hyperglycemic episode at morning monitoring times during the 7 days before injection was compared with the number of patients reporting hyperglycemic events at morning monitoring times during the first 7 days after injection by chi-square test. Insulinand noninsulin-dependent study subgroups were compared by analysis of variance repeated measures. The a level for all statistical hypothesis tests was set at DECEMBER 2012 Volume 35 Number 12 e1755

3 Table Patient Data Parameter Value Average age (range), y 57 (34-71) Average baseline blood glucose level (range), mg/dl Morning 125 (91-201) Bedtime 153 ( ) Average preinjection HbA1c a (range) 6.6 ( ) No. men/women 9/14 No. insulin dependent 11 No. noninsulin dependent Diagnosis, No. 12 Trigger finger 12 dequervain s 8 Wrist tendonitis 2 Carpal tunnel syndrome Abbreviation: HbA1c, hemoglobin A1c. a Preinjection HbA1c was available for 12 of 23 patients. 1 A Results Twenty-seven (50%) patients returned completed datasheets. Four missing preinjection values were excluded because this rendered paired statistical analysis impossible. The Table shows patient demographics. No complications occurred from the injections. No patient stated that he or she required additional medication to compensate for elevated blood glucose levels during the study period. Blood glucose levels before and after injection are shown in the Figure. The highest preinjection blood glucose level was 424 mg/dl for insulin-dependent patients and 214 mg/dl for noninsulin-dependent patients, both at bedtime. The highest postinjection blood glucose level was 518 mg/dl for insulin-dependent patients and 264 mg/dl for noninsulin-dependent patients, also at bedtime. Pre- and postinjection blood glucose levels were notably B Figure: Average morning blood glucose levels (mg/dl) before and after injection for each day. The asterisk indicates statistically significant increases from preinjection values observed for day 1 (P5.046) and day 5 morning values (P5.034). Average bedtime blood glucose levels (mg/dl) before and after injection for each day. The asterisk indicates statistically significant increases from preinjection values observed for day 6 bedtime values (P5.045). Abbreviations: insulin-dependent (ID); NID, noninsulin-dependent. variable. Average preinjection HbA1c was 6.6%61.3%. Average morning blood glucose levels remained above preinjection values until 7 days after injection. However, statistically significant differences from preinjection values were noted on days 1 and 5 after injection (14.2 and 9.7 mg/dl, respectively; P5.046 and.034, respectively). Average bedtime blood glucose levels remained above preinjection values until day 10 after injection, and day 6 after injection showed a statistically significant difference (32.7 mg/dl; P5.045). For postinjection days with nonsignificant differences, a post hoc power analysis found that this sample had 80% power to detect a difference of at least 20 to 46 mg/dl from morning preinjection values and 31 to 93 mg/dl from bedtime preinjection values as statistically significant, depending on the day. Sixty-eight percent of patients experienced at least 1 significant blood glucose increase postinjection (defined as an increase more than 2 times the standard deviation of preinjection values), and 53% of patients experienced at least 1 significant increase resulting in a blood glucose level higher than 180 mg/dl. The timing of these increases was inconsistent, ranging from e1756

4 Effect of Corticosteroid Injection on Blood Glucose Catalano et al immediately after injection to 4 weeks thereafter. On average, 16% (range, 0%- 31%) of patients experienced significant increases each day. Because morning data were more complete, proportions of patients experiencing at least 1 hyperglycemic event (higher than 180 mg/dl) at morning monitoring 1 week before and after injection were compared, and no significant difference was found (P5.64). However, the sample was underpowered for this comparison. Forty-three percent (10/23) of patients experienced more hyperglycemic events after injection than before injection. Of these, only 4 did not report a blood glucose level higher than 180 mg/dl before injection. These patients, with baseline values consistently lower than 180 mg/dl, reported a range of 0.5 to 3 hyperglycemic events per postinjection week. The remaining 6 patients who experienced an increased frequency of hyperglycemic events after injection (4 to 5.75 events per postinjection week) were hyperglycemic at least once per preinjection week, with recent HbA1c levels of 6.2%, 6.4%, 7.1%, 7.2%, and 7.5%, respectively (1 patient was missing data). Insulin-dependent patients experienced greater changes from baseline than did noninsulin-dependent patients. Significant differences were found between these groups for only the first 2 weeks postinjection (P5.012), and a trend existed toward significance for the entire study period (P5.055). Discussion Patients with diabetes mellitus have a propensity for inflammatory conditions affecting the hand and wrist. The higher incidence of these problems in patients with diabetes mellitus than in the general population poses a challenge for treatment because the local steroids used to treat these conditions are less effective in patients with diabetes mellitus. 1,4,7,15 In addition, patients are often concerned about how these locally administered drugs may affect glycemic control. Oral corticosteroids are widely known to have systemic effects on blood glucose control, but data conflict on the effect of intra-articular administration and limited data for extra-articular administration. 16,17 Habib and Miari 14 recently examined the effect of 40 mg of triamcinolone injected intra-articularly in the knee and found significant increases in blood glucose among patients with diabetes mellitus. Even et al 18 reported significant increases in blood glucose levels after epidural injection. In the only other study of the effect of smaller dose, extra-articular injections on blood glucose level, Wang and Hutchinson 10 reported a significant spike in blood glucose level, averaging a 73% increase that persisted for 5 days postinjection. The current authors tested a different steroid formulation, and their results differed in that no pronounced spike was observed, only a small average increase. The difference in results may be due to several factors. The study by Wang and Hutchinson 10 did not require patients to record blood glucose measurements before injection, instead asking patients for usual morning and afternoon measurements (approximately 100 and 120 mg/dl, respectively). The current patients preinjection morning averages ranged from 94 to 201 mg/dl, and average HbA1c was 6.6%61.3%. The American Association of Clinical Endocrinologists (AACE) recommended a target of 6.5% or less for most nonpregnant adults with diabetes mellitus. 19 The current study s sample appears to have marginally suboptimal control, with higher baseline values and variability, rendering most average postinjection increases statistically insignificant. It is likely that the cohort s suboptimal control represents a more realistic clinical situation. In addition, different steroid medications were administered in the studies: the current study used 10 mg of triamcinolone acetonide and Wang and Hutchinson 10 used 10 mg of methylprednisolone acetate. These doses are of equivalent potency; however, methylprednisolone acetate and triamcinolone acetonide have been shown to have different durations of action (7 to 84 vs 14 days, respectively) and serum half-lives (5 to 8 vs 3.2 to 6.4 days, respectively) Therefore, the current results may not be comparable. Many of the current study s patients reported significant yet isolated increases from preinjection values. By defining a significant individual increase as more than 2 times the standard deviation of the patient s baseline, 14 these substantial increases should be beyond normal fluctuations, indicating that the injection may have had an effect. No more than 31% of patients observed these increases on any day postinjection, which explains the lack of statistically significant increases in aggregate data except on postinjection days 1, 5, and 6. However, those statistically significant average increases (14.2, 9.7, and 32.7 mg/dl, respectively) are not likely to be clinically significant. The primary concern regarding increased blood glucose is whether it results in hyperglycemia, defined here as higher than 180 mg/dl. More than half of the current patients reported at least 1 such instance of blood glucose level higher than 180 mg/dl preinjection. Three patients with such events both before and after injection had a HbA1c higher than the AACE-recommended target of 6.5%. 19 The highest, 7.5%, indicates a mean plasma glucose level of mg/dl 23 ; thus, it would not be unexpected for such patients sugars to fluctuate to higher than 180 mg/dl. Regardless of baseline control, 43% of the current patients were more frequently hyperglycemic after injection, indicating the injection may have a hyperglycemic effect. Insulin-dependent patients experienced greater increases from preinjection values than did noninsulin-dependent patients, and the groups were significantly different. This difference was expected because, in general, insulin-dependent patients have more difficulty controlling blood glucose than do noninsulin-dependent patients. 24 DECEMBER 2012 Volume 35 Number 12 e1757

5 This study had several weaknesses. The follow-up rate was lower than expected. Patients were supplied with stamped, addressed envelopes and were called to remind them to return the forms after their monitoring periods, but many failed to do so. Although patients were asked to record 4 blood glucose levels per day, patients did not reliably report all 4 and returned study forms with missing data, limiting the power of the analysis for bedtime levels. The comparison of morning values was sufficiently powered, detecting a difference of 46 mg/dl or less as significant on any given postinjection day. Although the authors suspect that patients reported morning blood glucose levels were recorded when fasting, this can not be confirmed definitively. However, the AACE and the American Diabetes Association do not specifically recommend in clinical practice guidelines that patients self-monitor before breakfast. They recommend self-monitoring at least 2 to 3 times daily and before injecting insulin, but the optimal timing is unclear. 19,23 Because selfmonitoring fasting glucose is not specifically recommended for diabetes care, the authors do not believe this detracts from the results. Furthermore, this study attempted to follow patients with diabetes mellitus controlling their glucose levels in daily life, analyzing paired data (using each patient as his or her own control). Patients monitored morning and bedtime postinjection sugars as they normally would, so it is unlikely that postinjection numbers would be skewed by a change in monitoring routine. Conclusion Despite being less effective in patients with diabetes mellitus vs controls, steroid injections play an important role in the treatment of disease in patients with diabetes mellitus. The initial hypothesis of this study, that patients with diabetes mellitus may routinely become hyperglycemic after extra-articular corticosteroid injection, was not proven. The results indicate that local steroid injections for inflammatory processes of the hand and wrist in patients with diabetes mellitus are a clinically safe treatment option with respect to glycemic control. Despite an increased frequency of hyperglycemic events and isolated, significant individual increases, these sporadic events are unlikely to pose a significant risk to the patient. On administering local steroid injections of triamcinolone acetonide in the hand and wrist, physicians should advise patients that they may experience occasional increases in blood glucose during the month following injection, but such increases are usually isolated and are expected to resolve within a day. References 1. Nimigan AS, Ross DC, Gan BS. Steroid injections in the management of trigger fingers. Am J Phys Med Rehabil. 2006; 85(1): Hay EM, Paterson SM, Lewis M, Hosie G, Croft P. Pragmatic randomised controlled trial of local corticosteroid injection and naproxen for treatment of lateral epicondylitis of elbow in primary care. BMJ. 1999; 319(7215): Stahl S, Kaufman T. The efficacy of an injection of steroids for medial epicondylitis. A prospective study of sixty elbows. J Bone Joint Surg Am. 1997; 79(11): Stahl S, Kanter Y, Karnielli E. Outcome of trigger finger treatment in diabetes. J Diabetes Complications. 1997; 11(5): Koh S, Nakamura S, Hattori T, Hirata H. Trigger digits in diabetes: their incidence and characteristics. J Hand Surg Eur Vol. 2010; 35(4): Chammas M, Bousquet P, Renard E, Poirier JL, Jaffiol C, Allien Y. Dupuytren s disease, carpal tunnel syndrome, trigger finger, and diabetes mellitus. J Hand Surg Am. 1995; 20(1): Baumgarten KM, Gerlach D, Boyer MI. Corticosteroid injection in diabetic patients with trigger finger. A prospective, randomized, controlled double-blinded study. J Bone Joint Surg Am. 2007; 89(12): Griggs SM, Weiss AP, Lane LB, Schwenker C, Akelman E, Sachar K. Treatment of trigger finger in patients with diabetes mellitus. J Hand Surg Am. 1995; 20(5): Koehler BE, Urowitz MB, Killinger DW. The systemic effects of intra-articular corticosteroid. J Rheumatol. 1974; 1(1): Wang AA, Hutchinson DT. The effect of corticosteroid injection for trigger finger on blood glucose level in diabetic patients. J Hand Surg Am. 2006; 31(6): Derendorf H, Möllmann H, Grüner A, Haack D, Gyselby G. Pharmacokinetics and pharmacodynamics of glucorticoid suspensions after intra-articular administration. Clin Pharmacol Ther.1986; 39(3): Caldwell JR. Intra-articular corticosteroids: guide to selection and indications for use. Drugs. 1996; 52(4): Gearhart JG, Forbes RC. Initial management of the patient with newly diagnosed diabetes. Am Fam Physician. 1995; 51(8): , Habib GS, Miari W. The effect of intra-articular triamcinolone preparations on blood glucose levels in diabetic patients: a controlled study. J Clin Rheumatol. 2011; 17(6): Murphy D, Failla JM, Koniuch MP. Steroid versus placebo injection for trigger finger. J Hand Surg Am. 1995; 20(4): Younes M, Neffati F, Touzi M, et al. Systemic effects of epidural and intra-articular glucocorticoid injections in diabetic and non-diabetic patients. Joint Bone Spine. 2007; 74(5): Habib GS, Abu-Ahmad R. Lack of effect of corticosteroid injection at the shoulder joint on blood glucose levels in diabetic patients. Clin Rheumatol. 2007; 26(4): Even JL, Crosby CG, Sung Y, McGirt MJ, Devin CJ. Effects of epidural steroid injections on blood glucose levels in patients with diabetes mellitus. Spine (Phila Pa 1976). 2012; 37(1):E46-E Handelsman Y, Mechanick JI, Blonde L, et al. American Association of Clinical Endocrinologists Medical Guidelines for Clinical Practice for developing a diabetes mellitus comprehensive care plan. Endocr Pract. 2011; 17(suppl 2): Chrousos GP. Adrenocorticosteroids and adrenocortical antagonists. In: Katzung BG, Masters SB, Trevor AT, eds. Basic and Clinical Pharmacology. 11th ed. New York, NY: McGraw Hill Medical; 2009: Cole BJ, Schumacher HR Jr. Injectable corticosteroids in modern practice. J Am Acad Orthop Surg. 2005; 13(1): Skedros JG, Hunt KJ, Pitts TC. Variations in corticosteroid/anesthetic injections for painful shoulder conditions: comparisons among orthopaedic surgeons, rheumatologists, and physical medicine and primary-care physicians. BMC Musculoskelet Disord. 2007; 8: American Diabetes Association. Standards of medical care in diabetes Diabetes Care. 2012; 35(suppl 1):S11-S Davies M. The reality of glycaemic control in insulin treated diabetes: defining the clinical challenges. Int J Obes Relat Metab Disord. 2004; 28(suppl 2):S14-S22. e1758