56Medicine and Health / Rhode Island

�Naomi R. Kramer, MD

Obstructive Sleep Apnea: A Brief Overview For thePrimary Care Physician

Obstructive sleep apnea (OSA) af-fects 2% to 4% of middle-aged

adults.1 It is even more common inthe elderly. Although the primary carephysician has the opportunity to playa pivotal role in the detection of thisdisorder, most physicians have hadlittle or no formal training in OSA; andthey frequently underdiagnose the dis-order.2,3 The Walla Walla project2 dem-onstrated that with several educationalinterventions for physicians and pa-tients, the OSA detection rate signifi-cantly increased. This article willreview the typical presentation of OSA,diagnostic tests, and treatment optionsas well as follow-up once treatment isinitiated.

OSA is characterized by repetitive

partial or complete closure of the up-per airway during sleep despite contin-ued respiratory drive (Figure 1A4).The patient demonstrates increasinglynegative intrathoracic pressures as in-creasing ventilatory effort is generatedto attempt to open the airway. Theseevents are usually associated with abrief arousal and/or an oxygendesaturation and transienthypercapnea. These repetitive respira-tory-related arousals result in signifi-cant sleep fragmentation, which, incombination with the oxygendesaturation, result in subsequent day-time sleepiness and fatigue

An apnea refers to cessation of air-flow for more than 10 seconds. Anhypopnea is a reduction of airflow for

10 seconds. Both events areassociated with continued res-piratory effort. In contrast,central apneas have no airflowand no effort. The averagenumber of apneas andhypopneas per hour of sleep iscalled the apnea-hypopnea in-dex. The American Academyof Sleep Medicine (AASM)consensus statement5 suggestsit is not necessary to distinguishbetween apneas andhypopneas. Instead, the term“respiratory events” should beused to refer to both becausethey have similar pathophysi-ology and consequence. Morethan five obstructed respiratoryevents per hour of sleep areconsidered abnormal.

The AASM consensusstatement includes both symp-toms and sleep study data inthe definition of the obstruc-tive sleep apnea-hypopneasyndrome (OSAHS). OSAHSis defined as criteria A or B plusC. Criterion A: Excessive day-time sleepiness that is not bet-ter explained by other factors;

Criterion B: two or more of the fol-lowing that are not better explained byother factors: choking or gasping dur-ing sleep, recurrent awakenings fromsleep, unrefreshing sleep, daytime fa-tigue, impaired concentration; Crite-rion C: overnight monitoringdemonstrates five or more obstructedbreathing events per hour during sleep.

As these criteria suggest, fatigueand disrupted sleep are frequent symp-toms of OSA. (Table 1)

Several recent studies have sug-gested that certain key symptoms andassociations are useful in predictingwho will have OSA. Kump, et al6

found that the three symptoms mostpredictive of OSA are: Self-reportedsnoring, witnessed apnea, and sleepydriving. The positive predictive valueof their model was enhanced by includ-ing body mass index (BMI) and gen-der. Netzer, et al7 found a simpleself-administered patient questionnairehelped identify patients at high risk forOSA. Key symptoms include persis-tent symptoms (>3 to 4 times per week)in 2 or more questions regarding snor-ing, witnessed apnea or daytime sleepi-ness. Alternatively, persistentsymptoms in conjunction with hyper-tension or obesity were suggestive ofOSA. Simply adding questions regard-ing snoring, pauses, and daytime sleepi-ness to the primary care physician’sreview of systems will increase the like-lihood of detecting obstructive sleepapnea. If the patient has no reliablebed partner, the lack of a history ofsnoring, pauses, etc. has less signifi-cance. One may then need to rely onother symptoms and associated medi-cal conditions

The medical disorders most com-monly associated with OSA includehypertension and upper body obesity.Approximately 50% of patients withobstructive sleep apnea have hyperten-sion. Conversely, 25 to 30% of pa-tients from a hypertension clinic will

Table 1. Symptoms of ObstructiveSleep Apnea

SnoringWitnessed Apnea/gaspingChoking/shortness of breath arousalsRecurrent awakeningsNocturia (three times per night)Morning headacheExcessive daytime somnolenceAutomobile accident or near missDecreased memory/concentrationDepression/irritabilityEnuresisSexual dysfunction

Table 2. Risk Factors Associatedwith Obstructive Sleep Apnea

HypertensionUpper body obesityMale sexIncreasing ageAbnormal pharyngeal anatomyEnlarged tonsils and adenoidsRedundant pharyngeal tissueRetrognathiaNasal obstructionExcessive alcohol useUntreated hypothyroidismAcromegaly

57Vol. 85 No. 2 February 2002

have obstructive sleep apnea.8 A collarsize >17 in men, >16 in women, is as-sociated with an increased risk of OSA.Table 2 lists the predisposing or riskfactors commonly associated withOSA.

Although the exact mechanism isstill under investigation, sleep-relatedbreathing disorders have been associ-ated not only with hypertension, butalso with cardiovascular disease inde-pendent of shared risk factors such asobesity, age and gender.8 Sleep-relatedbreathing disorders (SRBDs) have alsobeen associated with an increased riskof stroke. It is not yet clear whetherthis association is due to the increasedstroke risk associated with hyperten-sion or whether SRBD is an indepen-dent risk factor.8 In either caseclinicians should have a high index ofsuspicion for OSA in patients with car-diovascular and cerebrovascular disease.

Although obstructive sleep apneadoes not occur more commonly in pa-tients with chronic obstructive pulmo-nary disease (COPD), patients withboth COPD and OSA (termed theoverlap syndrome) may present withhypercarbia, polycythemia, andcorpulmonale at an earlier point intheir disease (i.e. FEV1 > 1 liter) thanif they had COPD alone. A patientwith significant hypercarbia and anFEV > 1 liter should prompt a searchfor a concomitant disorder such asOSA or obesity hypoventilation.

Although a complete medical ex-amination is important in the evalua-tion of patients for sleep apnea, certainkey aspects of the examination shouldget special attention; specifically,weight (or BMI), blood pressure, nose,and oropharynx. It is important tonote whether the nasal passages arepatent or obstructed by polyps, swol-len turbinates, or boggy mucosa. Snor-ing and obstructive sleep apnea can becreated in normal non-apneic patientsby plugging the nose. Visualization ofthe palate, uvula, tonsils, and lateralpharyngeal walls is helpful in under-standing what factors may be affectingan individual’s breathing during sleep.

Once the clinical history sugges-tive of obstructive sleep apnea is ob-tained and physical examination

performed, it is appropriate to consideran overnight sleep study. Fullpolysomnography (16 channels ormore) yields the most information re-garding sleep architecture, respiratoryevents, associated arrhythmias, oxygensaturation, and concomitant sleep dis-orders. It currently remains the goldstandard to evaluate sleep disorders.Portable 4-channel studies are helpfulin confirming a diagnosis of OSA.However, more subtle respiratoryevents associated with sleep fragmen-tation rather than oxygen saturationmay be underestimated because sleepis not monitored. Similarly, portablerespiratory studies are inadequate toevaluate a general complaint of exces-sive sleepiness which may be due toother causes such as periodic limbmovement disorder. More complicatedmultichannel home monitors mayprove useful in the assessment of OSA.Lastly, night-to-night variability in thefrequency of respiratory events hasbeen described in patients with OSA.Therefore, even a single “negative”polysomnogram may not rule out OSAin cases of high clinical suspicion.

Treatment involves behavioral in-terventions in conjunction with medi-cal, dental or surgical interventions.Obesity, alcohol, tobacco, and sleepdeprivation have all been shown to ex-acerbate OSA. Therefore, behavioral

intervention should be aimed at weightloss, reducing evening alcohol con-sumption, tobacco cessation and avoid-ing sleep deprivation. Avoiding thesupine position in bed may also behelpful for some patients.

Positive airway pressure is themost effective intervention for OSA.9

This is most often delivered in the formof continuous positive airway pressure(CPAP) which applies positive pressurethroughout the upper airway prevent-ing collapse (Figure 1B). The patientwears a mask over the nose (or noseand mouth) which is attached via tub-ing to a “blower” and in-line humidi-fier. Usually a second polysomogramis performed to titrate CPAP to theoptimal pressure which eliminatessnoring and obstructive events. Oncethe best pressure is determined, CPAPis set up at the patient’s home by amedical equipment company withwhom the patient’s insurance companyhas a contract.

The array of new masks and thedevelopment of heated humidificationhave made CPAP much more user-friendly. If a patient feels uncomfort-able exhaling against CPAP, bi-levelpositive airway pressure may be tried.Patients with severe COPD andhypercarbia may feel more comfortablewith an expiratory pressure set 4 to 5cm lower than the inspiratory pressure

58Medicine and Health / Rhode Island

�

. . . clinicians shouldhave a high index ofsuspicion for OSA in

patients withcardiovascular and

cerebrovascular disease

rather than having a continuous pres-sure. CPAP is extremely effective formost patients. However, complianceis in the 50% to 70% range. This isnot significantly different from com-pliance with other pulmonary treat-ments.10 With the addition of newmasks, new pressure settings, andheated humidification, compliance willhopefully improve.

Dental appliances work by mov-ing the lower jaw and hence the tongueforward away from the palate and pos-terior wall of the pharynx. Eveloff etal11 found that it also elevates the pal-ate. The overall efficacy of a dentalappliance for mild to moderate OSAis approximately 60%. The better ap-pliances are adjustable so that the po-sition of the jaw may be adjustedaccording to tolerance and symptoms.A follow-up sleep study with the den-tal appliance in place is necessary todocument adequate control of OSA.Severe OSA is not likely to be con-trolled with a dental appliance alone.

Surgical options for OSA includetraditional uvulopalato-pharyngoplasty(UPPP) alone or in conjunction withprocedures to move the lower jaw for-ward.12 Tracheostomy is extremely ef-fective but rarely offered now becauseof its cosmetic effects and associatedcomplications and because CPAP is soeffective. The overall efficacy of UPPPis approximately 50%. Laseruvuloplasty (which removes less tissue)should be considered only for snoringnot OSA. For patients with OSA whodemonstrate narrowing posterior to thetongue, other procedures such as theinferior sagittal mandibular osteotomyand genioglossal advancement with hy-oid myotomy and suspension (GAHM)may be considered. Bi-maxillary man-dibular advancement (LeForte 1 proce-dure) has also been done for obstructivesleep apnea. The Stanford group12 hasstudied this extensively and has found ahigh success rate. However, lower suc-cess rates have been published fromother centers. This more invasive pro-cedure is usually reserved for those whofail UPPP or GAHM or have signifi-cant craniofacial abnormalities. Itshould be performed in centers experi-enced with this operation.

Because the overall efficacy of thestandard uvulopalatopharyngoplasty isnot high, all OSA patients who un-dergo surgery should have a follow-upsleep study approximately threemonths after surgery to reevaluate thedegree of residual sleep apnea. Theymay show some improvement in symp-toms and snoring with continued un-derlying significant sleep apnea.

The reason to treat OSA is to alle-viate symptoms and to decrease the as-sociated morbidity and mortality.Excessive sleepiness, impaired concen-tration, neurocognitive function andmood have all been shown to improvewith CPAP treatment of obstructivesleep apnea.13 Similarly, Findlay et al14

has shown that performance on drivingsimulator tests significantly improvesafter CPAP is initiated.

Reduction in blood pressure hasbeen demonstrated in hypertensivepatients following treatment of OSAwith CPAP. Similarly, mortality datafrom He, et al15 showed that for pa-tients with severe OSA, both CPAPand tracheostomy, but not UPPP wereassociated with improved survival com-pared to no treatment. Partinen, et al16

demonstrated that patients with OSAsuccessfully treated (by tracheostomy)had fewer cardiovascular events thanthose who were conservatively treated(weight loss recommendation).

Office follow-up for patients withsleep apnea following treatmentshould include questions again regard-ing residual snoring, witnessed pauses,excessive daytime sleepiness, sleepydriving, mood and neurocognitivefunction. If the patient is using CPAPor a dental appliance, it is importantto ascertain how many nights per week

and how many hours per night theyare using it. Nasal symptoms maylimit CPAP use. Therefore, specificquestions regarding nasal congestionand coryza need to be asked. Symp-toms may improve with use of topicalnasal steroids or oral antihistamines.The use of an in-line heated humidi-fier with CPAP significantly increasesmoisture delivery to the upper airwayand decreases nasal irritation andsymptoms. This is especially impor-tant in New England where indoorheating dries out the air.

Other questions that are impor-tant in follow-up for patients on CPAPregard comfort with their mask andskin integrity. Pressure points may bealleviated with small pads or cushions.A dry mouth in the morning may pointto air leaking through the mouth,which a chin strap may ameliorate.Mask and head straps do wear out andneed to be replaced periodically. Pa-tients’ use or tolerance of the machinemay decrease as the materials wear.They may be more comfortable withnew equipment. Lastly, if the patientredevelops symptoms of excessivesleepiness or snoring or has a signifi-cant weight change while on CPAP, itwould be reasonable to reevaluate theoptimal pressure with a repeat sleepstudy.

In summary, OSA is a commondisorder with significant morbidity andmortality. The morbidity relates to thesleepiness and associated automobileaccidents, associated cardiovascular dis-eases, and neurocognitive and person-ality changes. In the review of systems,the primary care physician can easilyscreen for this treatable, but often over-looked, disorder. In addition, if a pa-tient complains of fatigue or snoring,a more detailed history regarding othersleep symptoms is appropriate.

If symptoms and/or associated dis-orders suggest OSA, it is reasonable torefer the patient to a sleep center forfurther evaluation. Once treatment isinitiated and the patient stabilized, theprimary care physician can join thesubspecialist in screening for recurrenceof OSA symptoms and assessing com-pliance with treatment.

59Vol. 85 No. 2 February 2002

Welcome to Citizens Bank Private Banking Group

T H E P R I VAT E B A N K I N G G R O U Pspecializing in lending to Medical professionals, like you.

C O M M E R C I A L L E N D I N G S E R V I C E SOur experienced Relationship Managers provide supe-rior service and financial resources. The commerciallending services include:

• Working Capital Revolving Lines of Credit• Term Loans for equipment• Commercial Mortgages (owner occupied) • Loans to assist in professional practice start-ups

O T H E R B A N K S E R V I C E SOur clients work with a Relationship Manager who willprovide day-to-day services, create an integrated andpersonalized approach and coordinate a resource team,including:

• Retail Branch Banking• Cash Management Services• Trust Services• Online Banking and Online Purchasing

(BizMart)

S C H E D U L E A M E E T I N GWe offer competitive terms and extraordinary customerservice to help meet your commercial banking needs.To schedule a meeting, contact one of our RelationshipManagers or myself.

IN RHODE ISLAND & SOUTHEASTERNMASSACHUSETTS, OUR TEAM INCLUDES:

Susan Coolbaugh (401) 455-5321Karen Kempskie-Aquino (401) 455-5333Robert O’Neill (401) 282-4338

Anthony A. Botelho (401) 456-7308 Vice President/ManagerPrivate Banking Group

REFERENCES1. Young T, Palta M, Dempsey J, et al.

The occurrence of sleep-disorderedbreathing among middle-aged adults.NEJM 1992; 328:1230-5.

2. Kramer NR, Cook TE, Carlisle CC,et al. The role of the primary care phy-sician in recognizing obstructive sleepapnea. Arch Intern Med 1999;159:965-8.

3. Ball EM, Simon RD, Tall AA, et al.Diagnosis and treatment of sleep ap-nea within the community: The WallaWalla project. Arch Intern Med 1997;157:419-24.

4. Sleep Apnea: Diagnosis and Treatmentslide set; 1997 American Sleep Disor-ders Association.

5. Littner MR, Shepard JW. Recommen-dations for research into measurementand classification of sleep disorderedbreathing: gazing into the crystal ball.AASM Task Force. Sleep 1999;22:667-89.

6. Kump K, Whalen C, Tishler PV, et al.Assessment of the validity and utilityof a sleep questionnaire. Am J RespirCrit Care Med 1994; 150:735-41.

7. Netzer NC, Stoohs RA, Netzer CM,et al. Using the Berlin questionnaireto identify patients at risk for the sleep

apnea syndrome. Ann Intern Med1999;131:485-91.

8. Roux F, D’Ambrosio C, Mohsenin V.Sleep-related breathing disorders andcardiovascular disease. Am J Med2000;108:396-402.

9. National Center on Sleep DisordersResearch. Sleep Apnea: Is your pa-tient at risk? Bethesda, MD: NationalInstitutes of Health; 1995. NIH pub-lication 95-3808.

10. Dekker FW, Dieleman FE, KapteinAA, et al. Compliance with pulmo-nary medication in general practice.Eur Respir J 1993;6:886-90.

11. Eveloff SE, Rosenbert CL, CarlisleCC, et al. Efficacy of a Herbst man-dibular advancement device in ob-structive sleep apnea. Am J Respir CritCare Med 1994; 149:905-9.

12. Sher AE, Schechtman KB, Piccirillo JF.The efficacy of surgical modificationsof the upper airway in adults with ob-structive sleep apnea syndrome. Sleep1996; 19:156-77.

13. Lamphere J, Roehrs T, Wittig R, et al.Recovery of alertness after CPAP inapnea. Chest 1989; 96:1364-7.

14. Findley LJ, Levinson MP, Bonnie R.Driving performance and automobileaccidents in patients with sleep apnea.Clin Chest Med 1992;13:427-35.

15. He J, Kryger MH, Zorick FJ, et al.Mortality and apnea index in obstruc-tive sleep apnea: Experience in 385male patients. Chest 1988; 94::9-14.

16. Partinen M, Jamieson A,Guilleminault C. Long-term outcomefor obstructive sleep apnea syndromepatients. Chest 1988; 94:1200-4.

Naomi R. Kramer, MD, who is inprivate practice, is Clinical Assistant Pro-fessor of Medicine, Brown MedicalSchool, She formerly was Associate Di-rector, Sleep Disorders Center of LifespanHospitals.

CORRESPONDENCE

Naomi R. Kramer, MD300 Richmond St., Suite 203Providence, RI 02903phone: (401) 274-5716fax: (401) 272-2646

60Medicine and Health / Rhode Island

�Nicholas S. Ward, MD, and Aidan O’Brien, MD

End of Life Issues In the Critically Ill

The process of dying has changeddrastically in the last century. In

the past, doctors simply did all theycould for a patient. When their treat-ments failed, their patients died, almostalways in their homes. Currently, in theUnited States about 80% of people diein a healthcare facility (60% in acutecare facilities),1 despite the fact thatabout 90% of Americans polled say theywould wish to die at home.2 This dis-parity is caused by two factors. First,many people die while undergoing treat-ments meant to postpone death. Sec-ond, many families feel they are unableto care for a dying person or are un-comfortable having a loved one die athome. The net result is that most peoplewill die in a hospital, or other healthcarefacility, and most likely undergo highlevels of medical care. The RobertWood Johnson Foundation estimatesthat about 20% of Americans will diein an intensive care unit or be treated inan intensive care unit just prior to death.

Two conclusions can be drawn.First, a tremendous amount ofhealthcare is being delivered to dyingpatients. This has been reflected inseveral studies like that of Cher and co-workers in 1997 showing that a rela-tively large percentage of Medicareexpenditures goes to treat patients inthe last weeks of their lives.4 Second,doctors practicing in America todaymust learn skills not necessary in thepast. Doctors need to recognize pa-tients who are going to die despitemedical care and help decide which ofthe medical therapies are appropriateand which are not. They need to guidetheir patients through a maze of medi-cal options in an attempt to balancepreservation of life with quality of life,a daunting task. This paper will re-view some of the major medical, ethi-cal, and legal issues involved in theseend-of-life decisions. This paper willfocus only on patients who becomecritically ill acutely, not those withlong-term, progressive, terminal illness.

HOW ARE CRITICALLY ILL PA-TIENTS DYING IN HOSPITALS?

Most Americans today are dying inhealthcare facilities. Furthermore, stud-ies have shown that the vast majority ofthese deaths, about 75% or more, oc-cur only after the patient, or family, hasdecided to limit care.5-7 In two land-mark studies, Predergast and cowork-ers helped define just how patients diein ICUs. In their first study, they com-pared deaths in their ICU from two timeperiods, 1987-88 and 1992-93, to de-termine how often CPR was performedprior to death and how often limits wereplaced on care prior to death. Their datashowed that the incidence of CPR intheir ICU had declined from 49% to10% and that the incidence of limitingcare by withholding or withdrawingsome therapy had increased from 51%to 90% of all ICU deaths.8

To compare their data with therest of the country, the same investiga-tors did a large follow-up study, a yearlater. They collected data from over6,000 patient deaths occurring in 131ICUs in 38 states over a 6 month pe-riod and analyzed the data for the in-cidence of various limits of care. Theyfound that on the average only 25% ofpatients dying in ICUs got CPR priorto death. About 70% of patients hadsome restriction on care prior to deathand almost 50% of patients actuallyhad some medical therapy withheld orwithdrawn prior to death.6 It is impor-tant to note that these were deaths oc-curring in an ICU, a place establishedfor the most aggressive care.

The other data to emerge from thisstudy was the variability among ICUs.The incidence of patients dying with fullaggressive measures ranged from 4% inone ICU to 79% in another. Likewise,the incidence of withdrawing medicalsupport ranged from 0% to 79%, de-pending on the ICU. While the overallpractice of limiting care in ICUs is com-mon, there is tremendous variabilityfrom place to place in end-of-life care.

WHO DECIDES?Surrogate Decision making

The vast majority of people willdie with some limit of care in place,whether in or out of an ICU. Unfor-tunately, the patient rarely participatesin these decisions. Someone else gen-erally decides to limit a dying patient’scare 60 to 70% of the time.7,9 Onlyabout 15 to 20% of patients have anadvance directive at admission to hos-pital; and those advance directives areoften inadequate to handle anythingbut the most obvious treatment deci-sions. Therefore, the burden of diffi-cult decisions falls to a proxy (a legaldelegation) or surrogate (a non-legaldelegation). Most often, this is a fam-ily member.

The process of surrogate decisionmaking is fraught with problems.While most would agree that family orfriends are the best people to decidefor the patient, several studies haveshown that patients rarely discuss spe-cific treatment options with their prox-ies; and surrogate decisions correlatepoorly with what the patient wouldactually want done.10,11 Furthermore,a study by Hare et al. showed that sur-rogates often valued different aspectsof dying, such as pain and suffering,than the patients, who were more con-cerned with burdening families andamount of time left to live.10

Legal IssuesAll fifty states recognize the legal-

ity of a patient’s right to refuse medi-cal care although there remains somecontroversy and confusion about spe-cific issues. The legal issues involvedin proxy decision-making can be con-fusing. Perhaps because it is impos-sible to account for the many familyand social relationships that may be thesource of medical surrogates, moststates have few laws dealing with thisissue and have purposely kept the codesvague and malleable.12 Most states,including Rhode Island, will accept a

61Vol. 85 No. 2 February 2002

properly drafted written advance direc-tive as sufficient legal guidance to limitcare. Unfortunately, most advance di-rectives or living wills are too vague,using phrases such as “terminal illness”and “little chance of recovery” that aresubject to interpretation. COPD andcongestive heart failure may be consid-ered terminal illnesses by some peopleand not by others. In contrast, somepeople may consider diseases such asearly stage lung cancer not eminentlyterminal.

Nevertheless, these directives canhelp prevent futile or unwanted carewhen no other surrogate is available.More often they are useful in familydecision-making when an unconsciouspatient faces potentially futile care.The previously stated wishes of thepatient in an advance directive can as-suage guilt or uncertainty regardingend-of-life decisions. They can also behelpful when surrogates disagree as toa course of action. Since a surrogate,by definition, represents what the pa-tient would decide if able, the advancedirective can be a helpful guide.

Sometimes advance directives canspur discord - for instance, when thewritten directive differs from asurrogate’s decision. In most states in-cluding Rhode Island, the law recog-nizes a properly drafted and witnesseddirective as the legal opinion thatshould be followed; however, manyphysicians would be wary of ignoringthe requests of a living surrogate, espe-cially if it is a spouse or other close fam-ily member. In such situations,attempts should be made to build con-sensus among all parties prior to mak-ing any decision. Most state lawsregarding written advance directivesalso allow for some flexibility in thephysician’s obligation to follow them.For example, if a physician questionsthe validity of the directive or feelsethically unable to follow the directive,in most states the directive will not bebinding.

Predicting OutcomesA central problem complicating

end-of-life decisions is the difficulty ofpredicting outcomes in critically illpatients. The combination of multiple

coinciding medical problems and rap-idly changing clinical status can makethis a very difficult task. Essentiallythe physician has three tools: publishedoutcomes, severity scores, and personalexperience. All can be helpful yet allhave limitations.

Severity ScoresSeverity Scores have been available

for almost three decades. In most se-verity score algorithms data are col-lected during the first twenty-fourhours of admission and used to com-pile a score that, theoretically, predictsrisk of death during hospitalization.These scoring systems were developedby reviewing data from thousand ofICU patients and employing logisticalregression models to choose someimportant input variables. Other vari-ables were simply chosen based on pre-sumed clinical value. These scores werethen validated prospectively on pa-tients.

Unfortunately, there are severalproblems with these systems. First,these scoring systems make predictionsbased on hospital outcomes at the timeof their creation. As medical treat-ments improve, the scores need to beupdated. In the 1970s, for example,ARDS had a mortality approaching80%; thus the diagnosis might justifi-ably increase a patient’s severity score.Today ARDS has about a 40% mor-tality; thus a severity scoring systememploying the diagnosis of ARDS, oreven components of the diagnosis suchas hypoxemia, would need to be ad-justed. Some commercially availableproprietary severity scoring systemssuch as APACHE III® are updated andrevalidated on a regular basis to avoidthis problem but many widely in use

today, such as APACHE II, are basedon patient data collected as long as twodecades ago.

Also, most models derive theirpredictions from factors present at orshortly after admission to the ICU, anddo not provide updated mortality esti-mates as the patient’s conditionchanges. Furthermore, severity scoresoften give intermediate mortality esti-mates such as 60% instead of clear yesor no answers. Even these numbers aresubject to confidence intervals. Per-haps the most glaring problem of se-verity scores is that they say nothingabout morbidity, disability, or survivalafter hospitalization. These factors areoften just as important as risk of deathin making end-of-life decisions. A pa-tient may accept a 30% chance of sur-vival if it were followed by a highquality of life, while not accepting a70% chance of survival if it were likelyto entail a poor quality of life.

OUTCOMES RESEARCH

Many of the same problems en-countered with severity scores apply tooutcomes data. While published out-comes studies remain an essential toolfor helping clinicians predict a courseof illness, they suffer from two majorproblems.

First, the population studied for aparticular illness may not share thesame characteristics as your particularpatient. In a recent large multicenterclinical trial of a new therapy for sep-sis, the mortality in the control (un-treated) population was 31%.13 It isimportant to note, however, that thistrial excluded patients with renal fail-ure, liver failure, pancreatitis, AIDS,and variety of other co-morbid condi-tions, thus limiting the usefulness ofthese data for prognostic purposes.

Second, therapies can change andimprove rapidly. In a series of fourpublished studies by different authorsbetween 1981 to 2000 examining themortality of pneumocystis carinii pneu-monia in ICU patients, the mortalitydecreased from 86% to approximately50%.14-17 Similar changes in outcomeover time have been reported with avariety of other illnesses such as ARDSas treatments have improved.

�

Perhaps the most glaringproblem of severity scoresis that they say nothing

about morbidity,disability, or survivalafter hospitalization.

62Medicine and Health / Rhode Island

PATIENT AUTONOMY VS. MEDI-CAL PATERNALISM

A central problem to the end-of-life decision making-process is defin-ing the role of the physician. Usuallythe physician is a combination of edu-cator and advisor, but this is not alwaysthe case. In the past physicians weremore likely to dictate courses of actionor treatment plans for their patients, aconcept referred to as medical pater-nalism. In many parts of the world tothis day, medical decisions are madethis way, with little input from the pa-tient or family. In these cultures pa-tients are comfortable with this kindof decision-making. More recently inthe United States, the concept of pa-tient autonomy has dictated medicaldecision-making. In its extreme form,patient autonomy holds that thephysician’s role is to educate the patientabout the problem and offer plausibletreatment plans, with their risks andbenefits. The patient would then in-dependently choose a course of action.Many physicians use this model ofpractice today, or a variant of it, feel-ing that it empowers patients, freeingthem from physician bias.

In contrast to this philosophy,many physicians and patients feel thephysician is obliged to offer a recom-mend course of action. While the dis-crepancies outlined here may not be ofgreat significance in deciding whetherto choose one medication over another,they take on tremendous significancewhen the decision is life or death. Ul-timately, each physician must deter-mine the degree of involvement he orshe feels is warranted in end-of-lifedecisions.

CONCLUSION

In summary, the process of dyingin America is changing rapidly. Whilethe physician has always had an im-portant role in the dying process, thatrole has now changed. Today’s physi-cian must not only be adept at admin-istering comfort measures, he or shemust decide when to initiate thosemeasures over other therapies aimed at

restoring health. Because the dyingprocess now involves the healthcaresystem more and more, physicians needto have good end-of-life skills morethan ever. Failure to address these is-sues will result in patients getting morepotentially futile care at the expense oftheir own comfort and increasing coststo the healthcare system.

REFERENCES1. Field M, Cassel C. Approaching Death:

Improving Care at the End of Life.Washington DC: National AcademyPress, 1997.

2. The Gallup Organization. Knowledgeand Attitudes Related to Hospice Care[National Hospice Organization Sur-vey. Princeton, NJ: Gallup Organiza-tion, 1996.

3. Linde-ZwirbleW, Angus D, Griffen M,et al. ICU care at the end-of-life inAmerica: An epidemiologic study. CritCare Med 2000; 28(12)supplement:A34.

4. Cher DJ, Lenert LA. Method of Medi-care reimbursement and the rate ofpotentially ineffective care of criticallyill patients. JAMA 2001; 278:1001-7.

5. Kellogg FR, Crain M, Corwin J,Brickner PW. Life-sustaining interven-tions in frail elderly persons. Talkingabout choices. Arch Intern Med1992;152:2317-20.

6. Prendergast TJ, Claessens MT, LuceJM. A national survey of end-of-lifecare for critically ill patients. Am JRespir Crit Care Med 1998; 158:1163-7.

7. Faber-Langendoen K. A multi-institu-tional study of care given to patientsdying in hospitals. Ethical and prac-tice implications. Arch Intern Med1996;156:2130-6.

8. Prendergast TJ, Luce JM. Increasingincidence of withholding and with-drawal of life support from the criti-cally ill [see comments]. Am J RespirCrit Care Med 1997;155:15-20.

9. Lynn J, Teno JM, Phillips RS, et al.Perceptions by family members of thedying experience of older and seriouslyill patients. SUPPORT Investigators.Study to Understand Prognoses andPreferences for Outcomes and Risks ofTreatments [see comments]. Ann In-tern Med 1997;126:97-106.

10. Hare J, Pratt C, Nelson C. Agreementbetween patients and their self-selectedsurrogates on difficult medical deci-sions. Arch Intern Med1992;152:1049-54.

11. Sulmasy DP, Terry PB, Weisman CS, etal. The accuracy of substituted judgmentsin patients with terminal diagnoses. An-nals Intern Med 1998;128:621-9.

12. Sabatino CP. The legal and functionalstatus of the medical proxy: Sugges-tions for statutory reform. J Law, Med,Ethics 1999;27:52-68.

13. Bernard GR, Vincent JL, Laterre PF,et al. Efficacy and safety of recombi-nant human activated protein C forsevere sepsis. NEJM 2001;344:699-709.

14. Wachter RM, Luce JM, Safrin S, et al.Cost and outcome of intensive care forpatients with AIDS, Pneumocystiscarinii pneumonia, and severe respi-ratory failure. JAMA 1995;273:230-5.

15. Mansharamani NG, Garland R,Delaney D, Koziel H. Management andoutcome patterns for adultPneumocystis carinii pneumonia, 1985to 1995: comparison of HIV-associatedcases to other immunocompromisedstates. Chest 2000;118:704-11.

16. Curtis JR, Rubenfeld GD. Aggressivemedical care at the end of life; doescapitated reimbursement encouragethe right care for the wrong reason?[Editorial]. JAMA 1997;278:1025-6.

17. Bedos JP, Dumoulin JL, Gachot B, etal. Pneumocystis carinii pneumoniarequiring intensive care management:survival and prognostic study in 110patients with human immunodefi-ciency virus. Crit Care Med 1999;27:1109-15.

Nicholas S. Ward, MD, is AssociateDirector, Rhode Island Hospital Medi-cal Intensive Care Unit, and AssistantProfessor of Medicine, Brown MedicalSchool.

Aidan O’Brien, MD, is Staff Phy-sician at the Providence VA Medical Cen-ter and the RI Hospital, and AssistantProfessor of Medicine, Brown MedicalSchool.

CORRESPONDENCE:Nicholas S. Ward, MDDepartment of Pulmonary andCritical Care MedicineRhode Island Hospital593 Eddy St.Providence, RI 02903phone: (401) 444-8760fax: (401) 444-3002e-mail: Nicholas_Ward@brown.edu

63Vol. 85 No. 2 February 2002

IMAGES IN MEDICINEEdited by John Pezzullo, MD

Images in Medicine: We encourage submission to the Images in Medicine section from all medical disciplines. Image(s) should capture the essence of how a diagnosis isestablished, and include a brief discussion of the disease process. The manuscript should be less than 250 words and include one reference. The manuscript and one or twocropped 5 by 7 inch prints should be submitted with the author’s name, degree, institution and e-mail address to: John Pezzullo, MD, Department of Radiology, Rhode IslandHospital, 593 Eddy St., Providence, RI 02903. An electronic version of the text should be sent to the editor at jpezzullo@lifespan.org.

Retrotracheal Parathyroid Adenoma

A 70 year-old female had elevated serum calcium on routine biochemistry profile. Her parathyroid hormone level wassubsequently found to be in the 200-300 mg % range (normal up to 72mg %). As part of the diagnostic work-up, aparathyroid scan was performed using Technetium-99m sestamibi, which showed a persistent focus of abnormal increasedactivity posterior, inferior, and medial to the right lobe of the thyroid gland [Figure 1]. The neck ultrasound was normal.Computed tomography of the neck was performed [Figure 2], which demonstrated a 1 cm mass (arrow) posterior to theright aspect of the trachea, corresponding to the finding on the parathyroid scan. The mass was resected and proved to bea parathyroid adenoma.

Approximately 3% of parathyroid adenomas are located ectopically within the mediastinum. Preoperative localizationreduces the morbidity and surgical exploration time when the adenoma is in an ectopic location. Most studies show asensitivity for detection of parathyroid adenomas of 90% using technetium 99m-sestamibi, with less sensitivity for MRI,CT scan, or ultrasound. In this case, the retrotracheal position of the adenoma obscured its visualization on ultrasound.

Barry Julius, MD, is a third year resident in diagnostic imaging at Rhode Island Hospital.Jac Scheiner, MD, is a staff radiologist in the Division of Nuclear Medicine at Rhode Island Hospital.

REFERENCE

Perez-Monte JE, Brown ML, Shah AN, et al. Parathyroid adenoma: Accurate localization and detection with Tc-99m sestamibi SPECT.Radiol 1996;85-91.

CORRESPONDENCE:Jac Scheiner, MDDepartment of Diagnostic ImagingRhode Island Hospitalphone: (401) 444-5184e-mail: JScheiner@lifespan.org

Figure 1. Figure 2.

64Medicine and Health / Rhode Island

Update On Treatment For Congestive Heart Failure�Andrew Sucov, MD

BACKGROUNDCongestive Heart Failure (CHF) is a common diagnosis in

the United States, with approximately 1 million hospital admis-sions and 40,000 deaths yearly attributable to it.1 In Rhode Is-land, the impact of CHF is also large - approximately 3,500admissions and 80 deaths annually (personal communication fromRI Department of Health). In late 1994, the Agency for HealthCare Policy and Research (AHCPR - now AHRQ) released aguideline for management of CHF, which was updated in 1999.This review will predominantly focus on two treatment modali-ties - the use of ACE inhibitors and spironolactone. Other com-mon treatments will be summarized at the conclusion of thereview.

PATHOPHYSIOLOGYCHF is the end result of myocardial damage or overload,

usually as a result of atherosclerotic cardiovascular disease. Theheart is not capable of keeping up with the body’s demand foroxygenated blood, leading to neurohumoral activation through-out the body, most notably an increase in adrenergic tone andstimulation of the renin-angiotensin-aldosterone (RAA) system.These responses, when kept in balance, enable the heart to func-tion further along the pressure-volume (Starling) curve and main-tain cardiac output; but when they become out of balance, serveto put additional stress on the heart and overload the body withfluid. Chronic management seeks to rebalance the physiologicchanges and enable the heart to perform, without producing sys-temic side-effects, along with preservation or even improvementof cardiac function.

ACE INHIBITORSACE inhibitors have reproducibly been shown to reduce

mortality and reduce progression of disease, especially in patientswith higher New York Heart Association (NYHA) levels of dis-ease severity.1,2 As a result of their efficacy and safety, the guide-line and major textbooks recommend them as standard treatmentfor virtually all patients with CHF, especially those with systolicdysfunction (LVEF < 40%). ACE inhibitors function via twodifferent pathways - vasodilation and blocking renin-angiotensin-aldosterone. In acute management, ACE inhibitors function pri-marily as vasodilators, improving cardiac output.3 On a chronicbasis, their role is more attributable to local moderation of renin-angiotensin-aldosterone levels.2 These help limit vasoconstric-tion and water retention. Given their generally well toleratedstatus and clear impact on mortality, current recommendationswould suggest that these should be first line agents, used ahead ofdiuretics, in patients of any functional class. While patients maysymptomatically improve at low doses, higher doses have beenshown to reduce mortality and patients should be titrated to these

levels when possible (captopril 150 mg/d, enalapril and lisinopril20 mg/d).2

Up to 10% of patients may have contraindications to ACEinhibitors. A new class of agents, the angiotensin receptor blockers(ARB), seems to avoid the angioedema and cough side effects.While it would appear that these agents should have similar im-pacts on CHF morbidity and mortality as ACE inhibitors, theliterature to date does not support a mortality benefit in CHFpatients.4 Until literature supports a mortality benefit, the ARBshould remain second line. Another second-line alternative forpatients with ACE inhibitor contraindications is the combinationof hydralizine and nitrates.2

SPIRONOLACTONESpironolactone is an aldosterone antagonist and a weak di-

uretic on its own. In combination with either other diuretics orACE inhibitors its effect on volume status may be quite signifi-cant. The benefits of spironolactone are two-fold - it does nothave the same negative effects on electrolytes as the most com-monly used diuretics, and as aldosterone is an essential compo-nent of the neurohumoral response to CHF, use of spironolactonemakes mechanistic sense to combat the deterioration in functionand mortality.5 The major concern for increased use of spirono-lactone is on potassium levels, as both ACE inhibitors and spirono-lactone may elevate the levels. Close monitoring should accompanyany switch in diuretic medication. A recent report suggests thataddition of spironolactone to standard treatment (ACE inhibi-tors, beta blockers and diuretics) led to reduced mortality andhospitalization in patients with NYHA class III or IV CHF.6 Asthis is only a single well-performed study, its results can’t be seen asconclusive for all patients. Regardless, the original guidelines sug-gest using spironolactone in patients with NYHA class IV CHF,further supported by this study. Additional studies may extendthese results to patients of less severe dysfunction.

OTHER TREATMENT OPTIONSFor patients in NYHA class II or III failure, beta blocker

(beta-1 selective or mixed alpha and beta blocker) use is consid-ered to be first line, along with ACE inhibitors.2 They appear tohave a greater effect on mortality than ACE inhibitors, likely be-cause of effects on neurohumoral status, arrhythmia supressionand reversal of pathophysiologic cardiac remodeling (carvedilolmay also increase LVEF).7 Contraindications include advancedAV block, MNA class IV failure and significant reactive airwaydisease.

Traditional diuretics have powerful effects on volume status,but no evidence suggests a mortality benefit. There is a significantnegative effect on potassium and magnesium, which may predis-pose patients with CHF to arrhythmias. Routine use should be

65Vol. 85 No. 2 February 2002

limited to patients with fluid retention, and typically in combi-nation with ACE inhibitors and beta blockers.2

Digoxin use has declined in the past few decades. No evi-dence suggests a mortality benefit in CHF. Electrolyte imbal-ances and toxicity are significant concerns, especially whencombined with diuretics. Digoxin may be useful in patients whoare unresponsive to ACE inhibitors and beta blockers or thosewith atrial fibrillation requiring rate control.2

REFERENCES1. Braunwald E. Heart Failure in Harrison’s textbook of Internal Medi-

cine [on-line edition]. Chapter 232 [28 screens]. Available from: URL:http://www.harrisononline.com.

2. Consensus Recommendations for the management of chronic heartfailure. Am J Card 1999;83(2A):1A-38A.

3. Sacchetti A, Ramoska E, Moakes ME, et. al. Effect of ED managementon ICU use in acute pulmonary edema. Am J Emerg Med 1999;17:571-4.

4. Losartan compared with captopril reduced mortality but not renal func-tion in heart failure [abstract]. ACP J Club 1997;127:9. Abstract of:Pitt B, Segal R, Martinez FA, et al., on behalf of ELITE study investiga-tors. Randomized trial of losartan versus captopril in patients over 65with heart failure (Evaluation of Losartan in the Elderly Study, ELITE).Lancet 1997;349:747-52.

5. Soberman JE, Weber KT. Spironolactone in congestive heart failure.Curr Hypertension Rep 2000;2:451-6.

6. Pitt B, Zannad F, Remme WJ, et al., for the Randomized Aldactone Evaluation StudyInvestigators. The Effect of spironolactone on morbidity and mortality in patients withsevere heart failure. NEJM 1999;341:709-17.

7. Bristow MR. Beta adrenergic receptor blockade in chronic heart failure. Circulation2000;101:558-69.

Andrew Sucov, MD, is a Clinical Coordinator at RIQP, and emer-gency physician at Rhode Island Hospital and Miriam Hospital.

CORRESPONDENCE:Andrew Sucov, MDPhone: (401) 528-3200Fax: (401) 528-3210E-mail: asucov@lifespan.org6SOW-RI-HF-02-06

The analyses upon which this publication is based were performed un-der Contract Number 500-99-RI02, entitled “Utilization and QualityControl Peer Review Organization for the State of Rhode Island,” spon-sored by the Health Care Financing Administration, Department ofHealth and Human Services. The content of this publication does notnecessarily reflect the views or policies of the Department of Health andHuman Services, nor does mention of trade names, commercial prod-ucts, or organizations imply endorsement by the U.S. Government.

The author assumes full responsibility for the accuracy and com-pleteness of the ideas presented. This article is a direct result of theHealth Care Quality Improvement Program initiated by the Health CareFinancing Administration, which has encouraged identification of qualityimprovement projects derived from analysis of patterns of care, and there-fore required no special funding on the part of this Contractor. Ideasand contributions to the author concerning experience in engaging withissues presented are welcomed.

– A Physician’s Lexicon –

A Radical Perspective on WordsThe radiologist, enjoying the radiant sunshine of a glorious

spring morning, interrupted his lunch of irradiated radishes toexamine the forearm of a political radical thought to have sus-tained a limb fracture. One look at the X-rays, however, eradi-cated any doubt that the radius had indeed been fractured.

This contrived paragraph contains eight words, of widelydifferent meaning, each a descendant of the Latin word, ra-dix, meaning root.

Mathematicians preceded physicians in exploiting theword, radix [as well as its plural, radices, and its diminutive,radicle ]. They defined radix [or, in English, radius] as anystraight line connecting two points; more specifically, as ameasurement of any linear spoke between the center and itssurrounding circle. A radius then came to mean any extensionfrom some central point spreading [or radiating] out in alldirections. Early anatomists perceived the principal forearmbone, the radius, as a spoke extending from the trunk of thebody to its periphery. The neuroanatomists were also not shyin expropriating radix. The proximal nerve roots of the spinalcord are named the radicles; and inflammatory disease of thesestructures, radiculitis.

Language usage over the centuries corrupted the wordradix to the word, ray, confining its meaning to a beam oflight extending outward from a solitary source of illumina-tion. [But when physicists then demonstrated that there wererays other than those within the range of visible light, the

meaing broadened to embrace suchentities as X-rays and gamma rays.] Physicians trained in thediagnostic and therapeutic uses of these rays were called radi-ologists, and these emanations came to be known as radiations[and when intentionally generated, the process was called irra-diation].

A shiny new fabric was synthesized by chemists duringthe last century. Because it glistened, they called it rayon.

As science contrives new technologies, the belabored word,radix, was repeatedly incorporated into many new words suchas radio, radium, radioactive, radiobiology, radiopelvimetry,radectomy [the extraction of dental roots] and even radar [anacronym of Radio Detecting and Ranging].

Botanists, perhaps because they are more grounded inearthy reality, retained the original Latin meaning of radix ;and thus small plant roots are called radicles and a particularlypungent root-derived vegetable is called a radish.

To a mathematician, a radical is a numeral which modi-fies a numeric root. But to an earlier historian a radical was aperson who sought out the fundamental or root meaning ofthings. Gradually, though, a radical came to mean an extrem-ist, someone favoring extreme solutions to social problems. Theword has now taken on a negative connotation as when RobertFrost said: “I never dared be radical when young for fear itwould make me conservative when old.”

– Stanley M. Aronson, MD