case study: treating hypertension in patients with …...ophthalmology service requested a med-icine...

C A S E S T U D I E S

Case Study: Treating Hypertension in Patients With Diabetes

Evan M. Benjamin, MD, FACP

PresentationL.N. is a 49-year-old white woman witha history of type 2 diabetes, obesity,hypertension, and migraine headaches.The patient was diagnosed with type 2diabetes 9 years ago when she presentedwith mild polyuria and polydipsia. L.N.is 5'4" and has always been on the largeside, with her weight fluctuatingbetween 165 and 185 lb.

Initial treatment for her diabetes con-sisted of an oral sulfonylurea with therapid addition of metformin. Her dia-betes has been under fair control with amost recent hemoglobin A1c of 7.4%.

Hypertension was diagnosed 5years ago when blood pressure (BP)measured in the office was noted to beconsistently elevated in the range of160/90 mmHg on three occasions. L.N.was initially treated with lisinopril,starting at 10 mg daily and increasingto 20 mg daily, yet her BP control hasfluctuated.

One year ago, microalbuminuriawas detected on an annual urine screen,with 1,943 mg/dl of microalbumin iden-tified on a spot urine sample. L.N.comes into the office today for her usualfollow-up visit for diabetes. Physicalexamination reveals an obese womanwith a BP of 154/86 mmHg and a pulseof 78 bpm.

Questions1. What are the effects of controlling

BP in people with diabetes? 2. What is the target BP for patients

with diabetes and hypertension?3. Which antihypertensive agents are

recommended for patients withdiabetes?

137CLINICAL DIABETES • Volume 22, Number 3, 2004

CommentaryDiabetes mellitus is a major risk factorfor cardiovascular disease (CVD).Approximately two-thirds of peoplewith diabetes die from complications ofCVD. Nearly half of middle-aged peo-ple with diabetes have evidence of coro-nary artery disease (CAD), comparedwith only one-fourth of people withoutdiabetes in similar populations.

Patients with diabetes are prone to anumber of cardiovascular risk factorsbeyond hyperglycemia. These risk fac-tors, including hypertension, dyslipi-demia, and a sedentary lifestyle, are par-ticularly prevalent among patients withdiabetes. To reduce the mortality andmorbidity from CVD among patientswith diabetes, aggressive treatment ofglycemic control as well as other cardio-vascular risk factors must be initiated.

Studies that have compared anti-hypertensive treatment in patients withdiabetes versus placebo have shownreduced cardiovascular events. The Unit-ed Kingdom Prospective Diabetes Study(UKPDS), which followed patients withdiabetes for an average of 8.5 years,found that patients with tight BP control(< 150/< 85 mmHg) versus less tightcontrol (< 180/< 105 mmHg) had lowerrates of myocardial infarction (MI),stroke, and peripheral vascular events. Inthe UKPDS, each 10-mmHg decrease inmean systolic BP was associated with a12% reduction in risk for any complica-tion related to diabetes, a 15% reductionfor death related to diabetes, and an 11%reduction for MI. Another trial followedpatients for 2 years and compared calci-um-channel blockers and angiotensin-converting enzyme (ACE) inhibitors,

with or without hydrochlorothiazideagainst placebo and found a significantreduction in acute MI, congestive heartfailure, and sudden cardiac death in theintervention group compared to placebo.

The Hypertension Optimal Treat-ment (HOT) trial has shown that patientsassigned to lower BP targets haveimproved outcomes. In the HOT trial,patients who achieved a diastolic BP of< 80 mmHg benefited the most in termsof reduction of cardiovascular events.Other epidemiological studies haveshown that BPs > 120/70 mmHg areassociated with increased cardiovascularmorbidity and mortality in people withdiabetes. The American Diabetes Associ-ation has recommended a target BP goalof < 130/80 mmHg. Studies have shownthat there is no lower threshold value forBP and that the risk of morbidity andmortality will continue to decrease wellinto the normal range.

Many classes of drugs have beenused in numerous trials to treat patientswith hypertension. All classes of drugshave been shown to be superior to place-bo in terms of reducing morbidity andmortality. Often, numerous agents (threeor more) are needed to achieve specifictarget levels of BP. Use of almost anydrug therapy to reduce hypertension inpatients with diabetes has been shown tobe effective in decreasing cardiovascularrisk. Keeping in mind that numerousagents are often required to achieve thetarget level of BP control, recommend-ing specific agents becomes a not-so-simple task. The literature continues toevolve, and individual patient conditionsand preferences also must come intoplay.

Diuretics have been shown to havesynergistic effects with ACE inhibitors,and one could be added. Because L.N.has migraine headaches as well as dia-betic nephropathy, it may be necessary toindividualize her treatment. Adding a β-blocker to the ACE inhibitor will certain-ly help lower her BP and is associatedwith good evidence to reduce cardiovas-cular morbidity. The β-blocker may alsohelp to reduce the burden caused by hermigraine headaches. Because of the pres-ence of microalbuminuria, the combina-tion of ARBs and ACE inhibitors couldalso be considered to help reduce BP aswell as retard the progression of diabeticnephropathy. Overall, more aggressivetreatment to control L.N.’s hypertensionwill be necessary. Information obtainedfrom recent trials and emerging newpharmacological agents now make it eas-ier to achieve BP control targets.

Clinical Pearls• Hypertension is a risk factor for car-

diovascular complications of diabetes.• Clinical trials demonstrate that drug

therapy versus placebo will reduce car-diovascular events when treatingpatients with hypertension and diabetes.

• A target BP goal of < 130/80 mmHgis recommended.

• Pharmacological therapy needs to beindividualized to fit patients’ needs.

• ACE inhibitors, ARBs, diuretics, andβ-blockers have all been documentedto be effective pharmacological treat-ment.

• Combinations of drugs are often nec-essary to achieve target levels of BPcontrol.

• ACE inhibitors and ARBs are agentsbest suited to retard progression ofnephropathy.


While lowering BP by any meanswill help to reduce cardiovascular mor-bidity, there is evidence that may helpguide the selection of an antihyperten-sive regimen. The UKPDS showed nosignificant differences in outcomes fortreatment for hypertension using an ACEinhibitor or a β-blocker. In addition, bothACE inhibitors and angiotensin II recep-tor blockers (ARBs) have been shown toslow the development and progression ofdiabetic nephropathy. In the Heart Out-comes Prevention Evaluation (HOPE)trial, ACE inhibitors were found to havea favorable effect in reducing cardiovas-cular morbidity and mortality, whereasrecent trials have shown a renal protec-tive benefit from both ACE inhibitorsand ARBs. ACE inhibitors and β-block-ers seem to be better than dihydropyri-dine calcium-channel blockers to reduceMI and heart failure. However, trialsusing dihydropyridine calcium-channelblockers in combination with ACEinhibitors and β-blockers do not appearto show any increased morbidity or mor-tality in CVD, as has been implicated inthe past for dihydropyridine calcium-channel blockers alone. Recently, theAntihypertensive and Lipid-LoweringTreatment to Prevent Heart Attack Trial(ALLHAT) in high-risk hypertensivepatients, including those with diabetes,demonstrated that chlorthalidone, a thi-azide-type diuretic, was superior to anACE inhibitor, lisinopril, in preventingone or more forms of CVD.

L.N. is a typical patient with obesity,diabetes, and hypertension. Her BP con-trol can be improved. To achieve the tar-get BP goal of < 130/80 mmHg, it maybe necessary to maximize the dose of theACE inhibitor and to add a second andperhaps even a third agent.

Volume 22, Number 3, 2004 • CLINICAL DIABETES138

SUGGESTED READINGS

ALLHAT Officers and Coordinators for theALLHAT Collaborative Research Group: Majoroutcomes in high-risk hypertensive patients ran-domized to angiotensin-converting enzymeinhibitor or calcium channel–blocker vs. diuretic:the Antihypertensive and Lipid-Lowering Treat-ment to Prevent Heart Attack Trial. JAMA288:2981–2997, 2002

American Diabetes Association: Hyperten-sion management in adults with diabetes (Posi-tion Statement). Diabetes Care 27 (Suppl.1):S65–S67, 2004

Geiss LS: Mortality in NIDDM. In Diabetesin America. 2nd ed. Harris MI, Cowie CC, SternMP, Boyko EJ, Reiber GE, Bennett PH, Eds.Washington, D.C., U.S. Department of Healthand Human Services, National Institutes ofHealth, 1995 (NIH publ. no. 95-1468), p.233–258

Hansson L, Zanchetti A, Carruthers SG,Dahlof B, Elmfeldt D, Julius S, Menard J, RahnKH, Wedel H, Westerling S: Effects of intensiveblood pressure lowering and low dose aspirin inpatients with hypertension: principal results ofthe Hypertension Optimal Treatment (HOT) ran-domized trial. Lancet 351:1755–1762, 1998

Huang ES, Meigs JB, Singer DE: The effectof interventions to prevent CV disease in patientswith type 2 diabetes mellitus. Am J Med111:633–642, 2001

Tuomelehto J, Rastenyte D, Birkenhager WH,Thijs L, Antikainen R, Bulpitt CJ, Fletcher AE,Forette F, Goldhaber A, Palatini P, Sarti C, FagardR: Effects of calcium blockade in older patientswith diabetes and systolic hypertension. N Engl JMed 340:677–684, 1999

U.K. Prospective Diabetes Study Group: Effi-cacy of atenolol and captopril in reducing risk ofmacrovascular and microvascular complicationsin type 2 diabetes (UKPDS 39). BMJ317:713–720, 1998

Wingard DL: Heart disease and diabetes. InDiabetes in America. 2nd ed. Harris MI, CowieCC, Stern MP, Boyko EJ, Reiber GE, BennettPH, Eds. Washington, D.C., U.S. Department ofHealth and Human Services, National Institutesof Health, 1995 (NIH publ. no. 95-1468), p.429–448

Evan M. Benjamin, MD, FACP, is anassistant professor of medicine and VicePresident of Healthcare Quality atBaystate Medical Center in Springfield,Mass.


Case Study: Atropine Ophthalmic AdministrationUnmasking Undiagnosed Diabetic Gastroparesis

Roger Kenneth Eagan, MD, and Pninit Varol, MD

PresentationR.R. is a 62-year-old white man withglaucoma and long-standing type 2 dia-betes complicated by peripheral neu-ropathy and retinopathy. He presented tothe emergency room with persistent nau-sea and vomiting. The patient wasadmitted with presumed symptomaticglaucoma. Three months earlier, he hadundergone pars plana vitrectomy surgeryfor a vitreal hemorrhage secondary to adiabetic tractional retinal detachment.The patient had developed subsequentneovascular glaucoma and had beeninstructed to use his ophthalmic medica-tions to control symptoms.

Several weeks before his emergencyroom visit, he began to experience lefteye pain. The patient was seen by hisophthalmologist, who diagnosed increas-ing intraocular pressure (IOP). The oph-thalmologist intensified his regimen andencouraged the patient to carefully fol-low the provided regimen. Soon after,R.R. began to suffer from progressivenausea and vomiting.

At the time of presentation, thepatient had been unable to keep solids orliquids down for several days. He wasadmitted and treated with intravenousfluids and promethazine, then dischargedafter 24 hours with arrangements for sur-gery the following week. The followingday, he returned with ongoing intractablenausea and vomiting with opthalmalgia.He underwent a successful shunt place-ment to relieve his IOP, which relievedhis opthalmalgia. However, he continuedto have severe nausea and vomiting. Theophthalmology service requested a med-icine consult for further evaluation of thenausea and vomiting.


The internal medicine consultantfound R.R. to be in significant distresswith intractable vomiting. His vitalsigns showed a temperature of 98.6°F,heart rate 88 bpm, respiratory rate 14,and blood pressure of 189/82 mmHg.Per ophthalmology, the eye appearedwell with ongoing normal IOP. Heartand lungs were unremarkable. Hisabdominal exam was unremarkable.Neurological exam demonstrateddecreased sensation in the feet in astocking pattern with no other apprecia-ble defects. A work-up for commoncauses of intractable nausea and vomit-ing using laboratory and radiologicalevaluation was unremarkable.

The diagnosis of gastroparesis wasentertained. His atropine ophthalmicsolution was discontinued. The patient’ssymptoms improved such that he wasagain able to take food by mouth. Oph-thamology, however, felt that for thelong-term benefit of his eyes, it wasimperative that the patient be restartedon the atropine ophthalmic solution. Fol-lowing reinstitution of the ophthalmicatropine, his nausea and vomitingreturned.

A gastric emptying study using Tc-99m sulfur colloid was obtained. Itshowed gastric emptying delay of 43.9%(normal range 8–28%). To optimizesymptom management and maintain thenecessary ophthalmic regimen, metoclo-pramide and erythromycin were begunwith good symptomatic relief.

Epilogue. Upon further questioning,R.R. and his wife reported a gradualdecrease in his meal sizes and increasein meal frequency over the past year. Hemost likely had been self-managing his

progressive diabetic gastroparesis. Withthe addition of the anticholinergic med-ication, his underlying diabetic gastro-paresis became clinically apparent, lead-ing to his admission and subsequentwork-up and diagnosis.

R.R. was eventually taken off theatropine ophthalmic drops, but continuedto have mild symptoms of diabetic gas-troparesis. Therefore, he was continuedon metoclopramide with success.

Questions1. Can atropine ophthalmic solutions be

absorbed in clinically significantamounts?

2. Is systemic absorption of other oph-thalmic drugs known to be clinicallysignificant?

3. What is a reasonable approach to usewith patients on ophthalmic agents?

CommentaryPatients with diabetes are known todevelop autonomic regulatory problems.Because of this, they can be especiallysusceptible to medications that haveeffects on the autonomic nervous sys-tem. Oral preparations of β-blockers andtricyclic antidepressants have been welldescribed. However, we rarely think ofophthalmic agents in this light. It wouldmake intuitive sense that if systemicabsorption of ophthalmic agents canattain sufficient serological levels, therewould be an expected clinical effect.1,2

From our review of the basic sci-ence literature, we have determined thatthe atropine ophthalmic solutions arereadily absorbed from the nasal andgastric mucosa.3,4 One study that meas-ured biologically active atropine

anhydrase inhibitors, and cholinergicagonists.6

The following is our approach topatients on these ophthalmic medica-tions. • To minimize the systemic absorption

of all ophthalmic agents, patientsshould be directed to strictly instill theprescribed dosage only.

• They should be further instructed tocompress the lacrimal sac for 2–3minutes after installation of the eyedrops.

• Patients and clinicians need to beaware of the possible systemic sideeffects and be diligent in monitoringfor them. It is therefore recommendedthat, at the follow-up visits, a brief,focused history and physical examshould be performed targeted towardsthese side effects.

• If side effects are noted, patient educa-tion should be reviewed.

• If clinically significant symptomsremain, a dialogue among primarycare physician, sub-specialist, andpatient should be undertaken weighingthe risk and benefits of ongoingadministration.

Clinical Pearls• All medications with autonomic mod-

ulating properties should be givenwith caution to patients with diabetes.

• All ophthalmic agents should be mon-


(1-hyoscyamine) in sera following ocu-lar and intravenous administration notedsurprisingly similar concentrations.3

We performed a Medline literaturesearch and found only a few referencesto the clinical systemic effects that canensue from the ophthalmic use ofatropine.5 We were unable to find anycases of diabetic gastroparesisunmasked by atropine ophthalmic solu-tions. We also contacted the pharmaceu-tical makers of the atropine preparationand were informed that no similar eventhad been reported. It is our assertionthat given the above bioavailabilityinformation, undiagnosed clinical sideeffects are more prevalent than the liter-ature reflects.

One of the challenges of primarycare physicians is to monitor patients’medication lists. With our sub-specialistcolleagues adding medications appropri-ate to the conditions they are managing,sometimes side effects and interactionswill occur. The ophthalmic drops some-times are overlooked in this process.There can be significant systemicabsorption of these ophthalmic drops.The effects of β-blocker ophthalmicsolutions on the cardiovascular and res-piratory systems have been widely dis-cussed. However, all of the followingophthalmic agents have consistent datashowing systemic effects: prostaglandinanalogs, adrenergic agonists, carbonic


itored for symptoms of systemicabsorption.

• Proper patient education can helpminimize the amount of ophthalmicdrug absorbed systemically.

REFERENCES

1Stevens MJ: Diabetic autonomic neuropathy.In Up To Date. Rose BD, Ed. Wellesley, Mass.,Up To Date, 2004. Available online athttp://www.uptodate.com/index.asp

2Wehrmann T, Caspary WF: Diabetic auto-nomic neuropathy of the gastrointestinal tract. InUp To Date. Rose BD, Ed. Wellesley, Mass., UpTo Date, 2004. Available online athttp://www.uptodate.com/index.asp

3Kaila T, Korte JM, Saari KM: Systemicbioavailability of ocularly applied 1% atropineeye drops. Acta Ophthalmol Scand 77:193–196,1999

4Lahdes K, Kaila T, Huupponen R, SalminenL, Lisalo E: Systemic absorption of topicallyapplied ocular atropine. Clin Pharmacol Ther44:310–314, 1988

5Merlie GJ, Weitz H, Martin JH, McClay EF,Adler AG, Fellin FM, Libonati M: Cardiac dys-rhythmias associated with ophthalmic atropine.Arch Intern Med 146:45–47, 1986

6Alward W: Drug therapy: medical manage-ment of glaucoma. N Engl J Med339:1298–1307, 1998

Roger Kenneth Eagan, MD, was a chiefresident in internal medicine at IndianaUniversity in Indianapolis at the timethis article was written. Pninit Varol,MD, is an internal medicine and pedi-atric primary care physician for HealthNet, Inc., in Indianapolis, Ind.


Case Study: Potential Pitfalls of Using HemoglobinA1c as the Sole Measure of Glycemic Control

Huy A. Tran, FACE, FRACP, FRCPA; Diego Silva, MD; and Nikolai Petrovsky, FRACP, PhD

PresentationCase 1A 55-year-old man of Southeast Asiandescent presented with abnormal renalfunction for investigation. He had a 6-month history of gradual but progressivelethargy, tiredness, and poor concentra-tion. There was no history of inheritedor acquired kidney disease, and he hadbeen fit and healthy before recent onsetof symptoms. He had a strong familyhistory of type 2 diabetes with an affect-ed father, aunt, and brother. He was anonsmoker and consumed ~ 0.35 oz ofalcohol daily.

Clinically, he appeared well, with ablood pressure of 160/105 mmHg andregular pulse of 88 bpm. His weight was222 lb (110 kg) and height 5'8" (1.72 m)(BMI ~37 kg/m2). Fundoscopy showedevidence of early nonproliferative retinaldisease with macula sparing. His generalexamination was otherwise normal, withno evidence of nephromegaly, renalbruit, or microvascular disease.

Urinalysis revealed the presence ofproteinuria, confirmed to be 1.8 g/daywith no hematuria. He had a blood ureanitrogen (BUN) level of 99.7 mg/dl andcreatinine of 6.1 mg/dl with sodium of136 mmol/l and potassium of 5.4mmol/l. His fasting glucose level was122.5 mg/dl.

A follow-up oral glucose tolerancetest (OGTT) confirmed the presence ofimpaired fasting glycemia and impairedglucose tolerance, with a fasting glucoselevel of 117.1 mg/dl (6.5 mmol/l) and a2-hour post-OGTT glucose level of160.4 mg/dl (8.9 mmol/l). His hemoglo-bin A1c (A1C) levels (performed byhigh-performance liquid chromatogra-


phy) were 4.5 and 4.2% on two separateoccasions with abnormal chromatograms(Figure 1). Follow-up electrophoresisrevealed the presence of hemoglobin E.Further investigation of his renal abnor-mality, including biopsy, confirmed thepresence of diabetic nephropathy, withno other causes for chronic renal failurebeing apparent. On follow-up, his dia-betes control was excellent with dietaryand lifestyle management.

Case 2A 60-year-old woman with long-stand-ing obesity and hypertension presentedconcerned about the possibility of type 2diabetes given her strong family historyof the disease. Clinical examination wasunremarkable, with a blood pressure of

150/80 mmHg. Her fasting glucose was117.1 mg/dl (6.5 mmol/l) and A1C ontwo separate occasions was 5.5 and4.5%. The patient was reassured that shedid not have diabetes and was dischargedfrom the clinic without follow-up.

One year later, her fasting glucosewas 122.5 mg/dl (6.8 mmol/l). A follow-up OGTT revealed a fasting glucose of124.3 mg/dl (6.9 mmol/l) and a 2-hourglucose of 214.4 mg/dl (11.9 mmol/l),diagnostic of diabetes. Her A1C wasmeasured again and found to be 5.3%. Incontrast to her low A1C, her serum fruc-tosamine was elevated at 312 �mol/l,consistent with the presence of chronichyperglycemia.

A review of her hemoglobin chro-matogram (Figure 2) showed an abnor-

Figure 1. E pattern hemoglobinopathy interfering with the analysis of A1Cvalue.

culating red cells because of the glycosy-lation reaction between hemoglobin andcirculating glucose.1 In the presence ofexcess plasma glucose, the hemoglobinbeta-chain becomes increasingly glyco-sylated, making the A1C a useful indexof glycemic control. The importance ofA1C as an index of diabetes control wasreinforced by the Diabetes Control andComplications Trial (DCCT).2 This studydemonstrated a direct correlationbetween glycemic control as indicated byA1C and the likelihood of developinglong-term diabetes-related complications.

Because A1C is based on hemoglo-bin, both qualitative and quantitativevariations in hemoglobin can affect theA1C value. These factors need to beconsidered when interpreting A1Cresults and serve to limit the use of A1Cas a diagnostic test for diabetes.

Clinicians should also appreciate thedifferences in assay methods for A1C,which have relevance to the possibilityof interference. In the case of reducedtotal hemoglobin or increased turnoverof red blood cells (RBCs), the level ofA1C will be reduced even in the pres-ence of high ambient plasma glucose,thereby falsely lowering the A1C andlimiting its usefulness as a measure of


mal peak interfering with the isolation ofthe glycated hemoglobin. This was con-firmed with the repeat measurement. Herhemoglobin electrophoresis studyshowed an abnormal hemoglobin com-ponent. This was further confirmed to behemoglobin British Columbia.

Questions1. What is A1C, and how relevant is it

to the control of diabetes?2. What are the potential confounders

of A1C use in the assessment ofpatients with diabetes?

3. What are the available alternatives toA1C as markers of glycemic control?

4. Are A1C values diagnostic of dia-betes?

CommentaryA1C is the nonenzymatic glycated prod-uct of the hemoglobin beta-chain at thevaline terminal residue. The number 1cfollowing HbA represents the order inwhich this hemoglobin is detected onchromatography. Hence, other hemoglo-bin peaks are referred to as HbA1a1,HbA1a2, HbA1b, and so forth.

The A1C constitutes about 60–80%of total glycated hemoglobin. It is nor-mally present, albeit at low levels, in cir-


glycemia. In situations where the A1C islow, contrary to high day-to-day glucoselevels, attention should be paid to thehemoglobin concentration, the bloodsmear, and possibly hemolytic parame-ters to rule out the presence of anemia orhemolysis.

The other pathophysiologicalprocess that can affect the A1C value isthe structure of hemoglobin itself. Quali-tatively, any disorder that affects hemo-globin production, particularly the beta-chain, will affect the A1C results. In thecase of patients with beta-thalassemia,the absence of beta-hemoglobin chainsfor glycosylation invalidates the use ofA1C. In other hemoglobinopathies, thereis often the combination of abnormalhemoglobin plus associated excessiveintramedullary hemolysis. These, in turn,will lead to a falsely low A1C.

Case 1 illustrates this well. The abnor-mality in hemoglobin E is a point substi-tution of glutamine for lysine at position26 on the beta-chain (B26 glu → lys).Patients with this kind of hemoglobinopa-thy are likely to form glycated hemoglo-bin E1c instead of A1C, leading to a lowA1C level.3 In Case 2, the abnormality inthe hemoglobin British Columbia wasfound to be at codon number 101 [Glu(GAG) → Lys (AAG)] on the beta-chain,which interferes with glycosylation andhence falsely lowers the A1C level.4 Anysuspicion of a discordant A1C levelshould be followed up with a review ofthe hemoglobin chromatogram for anyabnormal peaks and hemoglobin elec-trophoresis, if indicated.

The immunoassay technique used isanother potential interference with themeasurement of A1C. This methodemploys various antibodies to detect theA1C fraction. If the antibodies recognizespecifically the N-terminus of the beta-chain, this assay will deliver falsely lowresults in situations where the number ofbeta-chains is either abnormal orreduced as demonstrated in our two cas-es. If not suspected, patients may bethought to have better glycemic controlthan is actually the case due to falselylow A1C results.

Figure 2. Abnormal peak of hemoglobin British Columbia on the hemoglobinchromatogram.

diagnosis of diabetes is at best contro-versial.8–10 The American Diabetes Asso-ciation does not recommend its use as adiagnostic tool and suggests it shouldonly be used for monitoring diabetes.11

The Australian and New Zealand posi-tion statement regarding new classifica-tion and criteria for diagnosis of diabetesmakes no reference to the use of A1C.12

Although it can be falsely low, there areother conditions that can lead to a falselyelevated A1C, including alcoholism,13

lead poisoning, opiate addiction,1 exces-sive use of salicylate (due to interfer-ence),3 and pregnancy. The increase isoften small and is not of clinical rele-vance, leaving diabetes in the majorityof cases as the primary cause of A1Celevation.

Clinical Pearls• A1C is an important marker of

glycemic control in patients with dia-betes.

• A1C is subjected to interference in thepresence of associated comorbiditiesincluding hemoglobinopathies, hemo-lysis, renal failure, and alcoholism.

• Its use in the diagnosis of diabetes iscontroversial and not recommended.

REFERENCES

1Goldstein DE, Little RR, Wiedmeyer HM,England JD, McKenzie EM: Glycated haemoglo-bin: methodologies and clinical applications. ClinChem 32:B64–B70, 1986

2DCCT Research Group: The effect of inten-sive treatment on the development and progres-sion of long term complications in the diabetescontrol in insulin dependent diabetes mellitus. NEngl J Med 329:977–986, 1993

3Kilpatrick ES: Problems in the assessment ofglycemic control in diabetes mellitus. DiabetMed 14:819–831, 1997

4Huisman THJ, Carver MFH, Efremov GD: A


In the presence of renal failure, theclinical utility of A1C is even more ques-tionable. The hemoglobin in renal dis-ease gets carbamylated due to condensa-tion of urea-derived cyanate with theN-terminal amino groups. This can sub-sequently read as a high A1C result whendetected by common methods such asion-exchange chromatography. The A1Clevel in renal failure thus represents abalance between the anemia associatedwith renal disease and hemoglobinadducts in renal failure. These two fac-tors often balance each other out with theeventual outcome being that the A1Cvalue is unchanged. Therefore, A1Ccould be employed usefully as a markerfor diabetes-related complications inpatients with uremia.5 The combinationof anemia and hemoglobinopathy result-ed in a falsely low A1C level in Case 1.

Other useful markers for diabetescontrol include total glycohemoglobin,which does not take into account thehemoglobin beta-chain and other blood-based glycated proteins such as fruc-tosamine. The latter is also readily avail-able in most laboratories and is reflectiveof mean glycemia but over a shorter timeof 15–30 days compared with 60–120days of A1C.6 While useful, these testshave not been as well validated as A1C.Although they have not been proven toreliably predict diabetes complications,extrapolation of the DCCT data wouldsuggest they should also be useful forthis purpose. Otherwise, in the presenceof abnormal hemoglobin, one is left withday-to-day variations in blood glucosereadings with which to monitor glycemiccontrol.

Despite previous reports7,8 advocat-ing it, the use of A1C as a tool for the


syllabus of human hemoglobin variants, 1996.Augusta, Ga., The Sickle Cell Anemia Founda-tion, 1996

5Lund L, Mourits-Andersen T, Sorensen PJ:Hemoglobin A1c and uremia. Clin Nephrol29:161–162, 1988

6Sacks DB, Bruns DE, Goldstein DE,Maclaren NK, McDonald JM, Parrott M: Guide-lines and recommendations for laboratory analy-sis in the diagnosis and management of diabetesmellitus. Clin Chem 48:436–472, 2002

7Peters AL, Davidson MB, Schriger DL, Has-sselblad V: A clinical approach for the diagnosisof diabetes mellitus: an analysis using glycosylat-ed haemoglobin levels. JAMA 276:1246–1252,1996

8Clutter WE: Meta-analysis: glycosylatedhaemoglobin levels are useful for diagnosing dia-betes. ACP Jrn Club 126:46, 1997

9Davidson MB, Schriger DL, Peters AL, Lor-ber B: Relationship between fasting plasma glu-cose and glycosylated hemoglobin: potential forfalse-positive diagnosis of type 2 diabetes usingnew diagnostic criteria. JAMA 281:1203–1210,1999

10Vinicor F: When is diabetes diabetes? JAMA281:1222–1224, 1999

11American Diabetes Association: Screeningfor diabetes (Position Statement). Diabetes Care25 (Suppl. 1):S21–S24, 2002

12http://www.mja.com.au/public/issues/apr19/colman/colman.html (last cited 12/5/2004)

13Hoberman HD, Chiodo SM: Elevation ofthe hemoglobin A1 fraction in alcoholism. Alco-hol Clin Exp Res 6:260–266, 1982

Huy A. Tran, FACE, FRACP, FRCPA, isdirector of the Division of Clinical Chem-istry of the Hunter Area Pathology Serviceat John Hunter Hospital in Newcastle,Australia. Diego Silva, MD, is a researchfellow at the Autoimmune Research Unitof the Australian National University inCanberra. Nikolai Petrovsky, FRACP,PhD, is a professor and director of theNational Health Sciences Unit of the Aus-tralian National University and a seniorendocrinologist in the Department ofEndocrinology and Diabetes of the Can-berra Hospital in Canberra.


Case Study: Skin Infection in a Diabetic PatientRelated to Contamination of an Insulin Bottle

Irma Gazeroglu, MD; Michael Borenstein, MD, PhD; and Maria P. Solano, MD

PresentationD.P. is a 59-year-old white Hispanicwoman with a 12-year history of type2 diabetes treated with a thiazolidin-dione and multiple daily injections ofinsulin. She presented to the outpa-tient clinic with a 10-week history ofpainful skin lesions on her abdomenthat had been increasing in size. Thelesions developed at the site of insulininjections. She was injecting in theabdomen, using a new needle eachtime. She had received a 14-daycourse of levofloxacin 7 weeks beforethe clinic visit and had been instructedto change the insulin bottles and touse her arms for injection. The skinlesions did not seem to improve, butshe did not developed new lesions.She denied fever or other constitution-al symptoms.

Her medical history was significantfor severe asthma requiring chronic oralsteroids and hypertension. Her medica-tions included rosiglitazone; irbesartan;prednisone, 20 mg daily; bronchodila-tors; and glargine and aspart insulins.Her glycemic control was poor, with ahemoglobin A1c result of 13.2%.

On physical examination, she hadCushingoid features and did not appearill. Her blood pressure was 120/60mmHg, heart rate 84 bpm, respiratoryrate 16 rpm, and temperature 98.4° F.On her abdomen, she had multiple ten-der, red, indurated, hemorrhagic crust-ed papules and nodules, 0.5–2 cm insize in the periumbilical region bilater-ally (Figure 1). There was no peripher-al edema, and there were no lesionselsewhere on her body.

Routine laboratory tests, including


leukocyte count with differential,platelets, electrolytes, creatinine, andliver enzymes, were within normalranges.

A skin biopsy was performed fromone of the nodules and was sent forhistopathology and culture.

Questions1. What is the microorganism involved

in this patient’s skin infection?2. How was the insulin bottle contami-

nated with the etiologic agent?

CommentaryThe biopsy demonstrated numerousacid fast bacilli in the inflamed dermis(Figure 2). Unfortunately, due to laberror, a culture was not performed.

In this patient, the insulin bottlewas the culprit. After she changed it,she did not develop new lesions. Uponfurther questioning, she admitted thatthere was water dripping in the refriger-ator where she kept the insulin bottle,as a possible explanation of how thebottle was contaminated with the envi-ronmental pathogen.

D.P. was treated with clar-ithromycin for 3 months with resolu-tion of the lesions and only mild resid-ual hyperpigmentation in the area.

Occasionally, mycobacteria are iso-lated from nodular skin lesions ofimmunosuppressed patients. Many cas-es are linked to injections, and diabeticpatients are at especially high risk. Theskin infections are usually due to M.abcessus, M. chelonea, M. fortuitum,and M. kansasii.

Nontuberculous mycobacteria growslowly. Even the rapid growers maytake 3–7 days to form visible colonieson media, whereas slow-growingmycobacteria take weeks or do notgrow at all. The slow growth compli-cates antibiotic susceptibility testing.Antibiotics may be degraded duringprolonged incubation.

These mycobacteria are notoriouslyresistant to most antituberculosis drugs.Debridement is best combined with twoor three antibiotic drugs. Most common-ly used antibiotics are clarithromycin,clofazimine, amikacin, rifabutin, andsulfonamide.

Figure 1. Pink nodules and pink,crusted, scaly papules coalescing intoplaques on the right mid-abdomen.

Figure 2. Multiple acid fast bacilli inthe dermis (original magnification 100�)

atypical mycobacteria in diabeticpatients who do not respond toantibiotic therapy for common skinpathogens.

• When suspected, it is imperative toinform the lab technician to usespecial media for atypicalmycobacterium isolation. Skinbiopsy and cultures are essential toguide the therapy.


It is important to consider the possi-bility of mycobacterial infection in cas-es that do not respond to standardantibiotic therapy. It is essential to per-form skin biopsy and cultures to evalu-ate the lesions in order to guide therapy.

Clinical Pearls• It is important to keep in mind the

rare but potential skin infection with


Irma Gazeroglu, MD, is a fellow at theUniversity of Miami, Division ofEndocrinology, Diabetes, and Metabo-lism; Michael Borenstein, MD,PhD, is aresident at the University of Miami,Department of Dermatology; and MariaP. Solano, MD, is an assistant professorof medicine at the University of Miami,Division of Endocrinology, Diabetes,and Metabolism in Miami, Fla.

case study: treating hypertension in patients with …...ophthalmology service requested a med-icine...

Documents