10 QUANTITATIVE ASPECTS OF

CLINICAL REASONING
LEE GOLDMAN

CHAPTER 10 QUANTITATIVE ASPECTS OF CLINICAL REASONING

The process of clinical reasoning is poorly understood but is based on factors such as experience and learning, inductive and deduetivc reasoning, interpretation of evidence (hat itself varies in reproducibility- and validity, and intuition that often is difficult to define. In an effort- to improve clinical reasoning, a number of attempts have been made to analyze quantitatively the many factors involved, including defining the cognitive approaches that clinicians apply to difficult problems, devising computerized decision support fcy stems that are designed to emulate certain features of decision making, and applying decision theory to understand how judgments should be reached. While each of these approaches has advanced the understanding of the diagnostic process, all have practical and/or theoretical problems that limit their direct applicability to the care of the individual patient^ Nevertheless, these preliminary attempts to apply the rigor and logic inherent in the quantitative method have provided significant insights intb the process by which clinical reasoning is accomplished, have identified ways in which the process may be improved, and have made it possible to\minimize certainlfeatures of the workup that arc not cost-effective. Thus, while clinical reasoning cannot be reduced to probabilities or numbers, attempts at quantitative analysis of the process may improve the ways in which Ihe problems of individual patientS arc approached and solved. In a simplified model, quantitative clinical reasoning includes five phases. The first consists of an investigation of the chief complaint through key questions that arc included in the history of the present illness (Tabic 10-1). These questions are supplemented by the past medical history and by a physical examination that emphasizes^ detailed investigation of potential key organ systems. In the second phase, the physician may select from an array of diagnostic tests, each with its own accuracy and usefulness for investigating the possibilities raised in the differential diagnosis. Since each lest has its costs, and some entail risk and discomfort as well, the physician must ask whefhci^the history and physical examination arc sufficiently diagnostic before ordering tests. Third, the clinical data must be integrated with test results to estimate the likelihood of conditions in the differential diagnosis. Fourth, the comparative risks and benefits of further diagnostic and therapeutic options must be weighed to reach a recommendation for the patient. In the fifth and final phase, this recommendation is presented to the patient, and after appropriate discussion of the options, a therapeutic plan is initiated. Each of the five steps in this simplified model of the clinical reasoning proccss can he analyzed individually. HISTORY AND PHYSICAL EXAMINATION It originally had been assumed that physicians begin investigating a patients chief complaint by obtaining a comprehensive history, which includes many, if not most, of the questions included in a full review of systems, and by performing an allinclusive physical examination. However, experienced clinicians begin to form hypotheses based on the chief complaint

a systematic, thorough, and complete history of the present illness, past medical history, review of systems, family history, social history, and physical examination. For example, if a patient presents with abdominal pain, the physician should gather information regarding its location and quality as well as the factors that precipitate and/or relieve it. The physician then asks questions relating to the diagnoses that may be suspected based cm the response to the initial questions. If the pain is suggestive of pancreatitis, the clinician would ask about alcohol intake, the use of thiazide diuretics or glucocorticosteroids, symptoms suggestive of concomitant gallbladder- disease, a family history of pancreatitis, and questions aimed at uncovering the possibility of a posterior penetrating ulcer. Alternatively, if the discomfort seems more typical of reflux esophagitis; a different sequence of questions would be- triggered. The use of iterative hypothesis testing encourages the physician to elicit detailed information in high-yield areas, without forgoing a systematic and thorough approach to the patient. Findings on the history and physical examination should influence each other. The history focuses the physical examination on certain organs, and findings on physical examination should encourage more detailed review of certain systems. As physicians procccd through this reasoning proccss with both the history and physical examination, a variety of issues may influence the accuracy of the decision-making process. First is the potential for some historical information or physical findings to be poorly reproducible, either because the patients responses vary or because different physicians clicit information differently or vary in the way they interpret, the answers. The careful use of clear and, when possible, precise questions can increase the reproducibility and validity of the medical history but still cannot eliminate all variability. When assessing the reproducibility of findings on the physical examination, two observers frequently agree that an uncommon abnormality such as an enlarged spleen is not present but agree less often when one of them thinks that it is present in a patient in whom it would not usually be expected. This principle can best^be demonstrated by understanding that some agreement always occurs by chance, and the likelihood of chance agreement is higher if the finding is either very common or very uncommon. For example, if two physicians each consider 90 percent of patients to be abnormal in some manner, such as having a systolic heart murmur, they will agree 81 percent of the time by chancc alone. In some studies of the reproducibility of common signs and symptoms, such as an enlarged liver, actual agreement rates have hot been substantially better than chance. Disagreement rates may ho reduced by emphasizing physical examination skills during medical (mining, by looking for other correlative physical findings, and by learning how physical findings correlate with the results of diagnostic tests. Therefore, when a clinician notes an unexpected and somewhat subjective abnormality lor which Ihcrc may be a'high rate of interobservcr disagreement, such as an unexpectedly enlarged spleen, other abnormalities that may often be associated with it, such us hepatomegaly of lyt|>hailcnoputhy.

43

TABLE 10-1 Phases of clinical reasoning and decision making

/ Investijaiion of the complaint by means of clinical cxiiniiiiatiou (history and physical examination) | 2 Ordering of diagnostic lesls, each with its own intrinsic accnnicy and usefulness J Integration ol clinical findings will) lest results to assess diagnostic probabilities I Weij'liinj', o| comparative tisks and benefits ol alternative courses ol action if: Determination of patient's preferences and development of a therapeutic plan and on Ihe responses to

initial questioning, and they ask further questions in a sequence that allows them to evaluate the initial hypotheses and, if necessary, shorten or amend the list of possibilities. Only a limited number\of diagnostic hypotheses can be entertained at any one time, and information is used to build a case for or against the most likely. In such a way, high-priority questions are selected from the almost limitless number that might be asked, and these specific questions arc incorporated into the history of the present illness. Often, a key response, such as a history of mclena, will be selected, a list of potential explanations for it will be formulated, and this list will then be trimmed, based on the response to more probing questions, so that a principal diagnosis can be selected and then tested. This process, termed iterative hypothesis testing, is an efficient approach to diagnosis and is preferable to attempts to gather every conceivable piece of information prior to formulating a differential diagnosis. Advocacy of iterative hypothesis testing does not argue against the need for

should

be

sought to

incrca.sc

(he likelihood that the

CHAPTER 10 QUANTITATIVE ASPECTS OF CLINICAL REASONING

spleen

would he

expected to be abnormal. In some situations, ordering a diagnostic test, such as an abdominal* ultrasound, to assess the finding more objectively should be considered if the test is sufficiently reliable. These comments about the factors that limit the reproducibility and validity of the medical history and physical examination do not denigrate their critical importance in clinical reasoning. Rather, they emphasize that care and diligence in the application of these skills arc necessary. For example, careful auscultation of the heart during various bedside maneuvers (see Chap. 188) has been shown to be remarkably accurate in determining the cause of systolic murmurs. When physicians use the history and the physical examination to arrive at a diagnosis, they are rarely certain of it. Therefore, it would be better to assess the likelihood of the diagnosis in terms of probabilities. All too frequently this probability is not expressed as an actual percentage but rather in such terms as nearly always, commonly,* sometimes, or rarely. Since different physicians may assign different probabilities to the same terms, these imprcci.se words frequently lead to major misunderstandings among physicians or between the physician and the patient. Physicians should be as rigorous and quantitative as possible in their assessments, anil when feasible, a quantitative expression of probability should be used. Fpr example, rather than saying that it is unlikely that a radiographic pattern is indicative of a carcinoma of the colon, it would be preferable, if possible, to provide a more precise indication of the probability of carcinoma with this radiographic pattern. A 10 to 15 percent probability of carcinoma may be interpreted as unlikely but from a clinicaLperspccti ve .usual ly warrants further evaluation because of the serious consequences of missing a potentially resectable tumor. Although such quantitative estimates would be desirable, they usually are not available in practice. Even experienced physicians often ard unable to estimate accurately the likelihood of particular conditions. There is a tendency to overestimate the likelihood of relatively uncommon conditions, and physicians are especially poor at quantifying probabilities that are very high or very low!7 For example, a physician may not know whether the probability of bacterial meningitis or of another disease that could be diagnosed by a lumbar puncture in a patient with a severe hcadachc is I in 20 or 1 in 2000. In both situations, the probability is low, but the decision as to whether a lumbar puncture should be performed may depend on this estimate. As was emphasized in Chap. 1, the history and physical examination have other important purposes. They allow the physician to evaluate the emotional status of the patient and td understand how the present problems fit into the context of the patients social and family life, and they encourage the development in the patient of confidence in the physician, which is so necessary for reaching an agreement on the coming plan of action. DIAGNOSTIC TESTS: INDICATIONS, ACCURACY, AND USEFULNESS A diagnostic test should be ordered for specified clinical indications, be sufficiently accurate to be efficacious for such indications, and be the least expensive and/or risky of the available efficacious tests. No diagnostic test is totally accurate, and physicians often have^difficulty interpreting test results# It is therefore critical to understand several commonly used terms in test analysis and epidemiology, including prevalence, sensitivity, specificity, positive predictive value, and negative predictive value (Table 10-2). Although reports of the accuracies of diagnostic tcsts^re commonly expressed initcrms of positive and negative prcdiciivt^valucs, these calcutateJ values are dependent on the prevalence of the disease in the population being studied (Table 10-3). A test with a particular sensitivity and specificity has different positive and negative predictive values when used in groups of patients that have different prevalences of disease. For example, a mildly abnormal alkaline phosphatase level in a young adult with a known lymphoma suggests hepatic involvement by the tumor (i.e., it is likely to be a true positive), while the same alkaline phosphatase level as part of a routine screening

TAIlLK 10*2 Definitions of commonly used terms In epidemiology and decision making Disease state

43

Test result Positive Negative Prevalence (prior probability) Sensitivity Specificity False-negative rate Pulse positive rale Positive Negative prod id i ve predictivc value value
Overall accuracy

Present

Absent

a (true positive) h (false positive) c (false negative) d (true negative) cc)!(a + d) + =b all = (o + + patients with the disease/all patients
tested = true-positive test results/all patients with (he disease = true-negative test = d/(b + d) results/all patients without the disease false-negative test rcsul(s/all patient* = c/(a += c) with (he disease

a/(a + c)

(csl results/all hl{h 1 </) falfcc-niHkiiivc patients without the disease test results/all positive a/(a += b)true-positive test results

d/(c + d) true-negative test results/all patients

with negative results + d) =+true-positive + true- negative test = (a +c d)/(a +b results/ all tests

battery of blood tests in an asymptomatic person of the same age is unlikely to be due to tumor (i.e., in this setting it is more likely to be a false positive). Although Ihc sensitivity and specificity of a test do not depend on the prevalence (or percentage of patients being tested who have the disease), they do depend on the spectrum of patients inf whom the test is being evaluated. For example, measurement of a prostatc- spccific antigen for diagnosing carcinoma of the prostate (sec Chap. 323) will appear to have a nearly perfect sensitivity and specificity if the diseased population has a palpably prostate noclulc and an elevated scrunl acid phosphatase level while the nondiscascd population is composed of normal medicalstudents. If, however, without changing the prevalence of disease in the population being tested, the spectrum of the diseased and nondiseased patients is altered by including oatients with other ^characteristics (tre., if the*population of patients 'with carcinoma of the prostate were composed principally of those without palpable nodules and with stage 1 disease, while the population without carcinoma of the prostate included elderly men with marked benign prostate hypertrophy), the sensitivity and specificity of the test would change dramatically. In the latter situation, the sensitivity and specificity of the prostate-specific antigen are not only lower than in the first example, because the spectrum of diseased and nondiscased patients has been changed, but more important, they may be so low that the test is of limited clinical value. This example also demonstrates the methodologic problems encountered when applying data from one study to a different type of patient or when pooling data fronustudies of different subsets of patients. In some situations, uncertainty about the sensitivity and specificity of the test in (he type of patient being assessed may limit its clinical value. Since the physician rarely knows (or enn know) the population on which every test that is ordered has been standardized, the results provide information that is far less decisive than usually thought. Furthermore, it may be quite difficult to distinguish random laboratory errors from test results that might be falsely positive or negative bccausc of coexistence of a process that can affcct the test, such as the finding of an elevated level of CK in a patient who has undergone strenuous exercise and is being evaluated for chest pain. Uccausc no single value or cutoff point of an individual test can be expected to have both a perfect sensitivity and a perfect specificity.

apericardium and central tendon of the diaphragm, is felt characteristically at the tifem-lhc-shoulder- the adjoining tfapezius ridge, and ihjg. neck involvement of the more lateral part of the diaphragmatic pleura, supplied by branches from the si&jJ] to ninth intercostal nerves, causes pain not only in the anterior part of the chest but also in the upper pail of the abdomen or corresponding region of the buck, sometimesjjimulating the pain of acute cholecystitis or pancreatitis* Pericardial puin commonly has a pleuritic component; i.e., it is related to respiratory movements and aggravated by cough and/or deep inspiration, bocause of pleural irritntion. It is sometimes brought on by swallowing, because the esophagus lies just behind the posterior portion of the heart, and is often altciyd^by a change of body.position, becoming_sharpcr and more lefraded in the supine position and f reduced when the patient sitsjbpright, leaning forward^It is frequently ^iefcrfctf"tcrthe'neelc and lasts^onger than the pain ofangina pectoris. In some patients, however, pericardial pain may be described as a steady substemal discomfort .mat.can mimic the pain of acute myocardial infarction. The mechanism of this steady substemal pain is not certain, but it may arise from marked inflammation of the relatively insensitive inner parietal surface of the pericardium or from irritated afferent cardiac nerve fibers lying irvthe periadventitial layers of the superficial coronary arteries. Occasionally, both pleuritic and steady pain may be present simultaneously. Patients with marked right ventricular hypertension may have exertional pain which is quite similar to that of angina. This discomfort probably results from relative ischemia of the right ventricle brought about by the increased oxygen needs and by the elevated intramural resistance, with reduction of the normally large systolic pressure gradient which perfuses this chamber. The pain due to acute dissection of the aorta (Chap. 210) or to an expanding aortic aneurysm results from stimulation of nerve endings in the adventitia. The pain usually begins abruptly, reaches an extremely severe peak rapidly, is felt in the center of the chest and/or in the back depending on the site of the dissection, lasts for hours, and requires unusually large amounts of analgesics for relief. Patients commonly describe a true pain rather that the vague discomfort that is sometimes described with myocardial ischemia. The pain is not aggravated by changes in position or respiration The pain resulting from pulmonary embolism (Chap. 226) may resemble that of acute myocardial infarction, and in massive embolism it is located substcrnally. In patients with smaller emboli, the pain is loeated more laterally, is plcuritic in nature, and may be associated with hemoptysis.

CHAPTER 10 QUANTITATIVE ASPECTS OF CLINICAL REASONING

The presence of an abdominal disorder, such as a hiatus hernia or a duodenal ulcer, docs not constitute proof that the patient's chest pain is related to it. Such disorders are frequently asymptomatic and arc not at all uncommon in patients who also have angina pectoris. Musculoskeletal pain The costochondral and chondrostemal articulations are the most common sites of anterior chest pain Objective signs in the form of swelling (Tietzes syndrome), redness, and heat are rare, but sharply localized tenderness is common. The pain may be darting and last for only a few seconds or may be a dull ache enduring for hours or days. An associated feeling of lightness due to muscle spasm (see below) is frequent.

43

Pressure on die chondrostemal and costochondral junctions and on the pectoralis muscles is an essential part of the examination of every patient with chest paint and will reproduce the pain arising from these
tissues. A large percentage of patients with costochondral pain, especially those who also have minor and innocent T-wave alterations, are erroneously labeled as having coronary disease. Pain secondary to subacromial bursitis, biceps tendonitis, and arthritis of the shoulder and spine may be precipitated by motion but not by general exertion. Pain arising in the chest wall or upper extremity may develop as a result of muscle or ligament strains brought on by unaccustomed exercise and felt in the costochondral or chondrostemal junctions or in the chest wall muscles. Other causes arc osteoarthritis of the dorsal or thoracic spine and ruptured cenicml disk disease. Pain in the left upper extremity and prccordium may be due to compression of portions of the brachial plexus by a cervical rib or by spasm and shortening of the scalenus anticus muscle because of high fixation of the ribs and sternum. Deep breathing, turning or twisting of the chest, and movements of the shoulder girdle and arm may elicit and duplicate the pain of which the patient complains. The pain may be very brief* lasting only a few seconds, or aching and persist for hours. The duration is therefore likely to be cither longer or shorter than untreated angina pectoris, which usually lasts for only a few minutes.

Pleural pain from fibrinous pleurisy or any pneumonic process is very common. It generally results from stretching of inflamed parietal pleura and is similar in character to the pleural pain of pericarditis (see above). Pneumothorax and tumors involving the pleural space also may irritate the parietal pleura and cause pleural pain; the latter is sharp, knifejikc, superficial in quality, and its aggravatiortby each breath and by coughing distinguishes it Irom the deep, dull, relatively steady pain of myocardial ischemia. Substemal discomfort also frequently occurs in the presence of tracheobronchitis; it is commonly described as a burning sensation accentuated by coughing.
The pain of mediastinal emphysema (Chap. 228) may be intense and sharp and may radiate from the substemal region to the shoulders; often a distinct crepitus is heard. The pain associated with mediastinitis and mediastinal tumors usually resembles that of plcuritis but is more likely to be maximal in the substemal region, and the; associated feeling of constriction or oppression may cause confusion with myocardial infarction. The several abdominal disorders which may at times mimic anginal pain may usually be suspected from the history, lisophugctil pain commonly prcscnt^a* a deep thoracic burning discomfort, which ivthe hallmark ol acidinduced pain. Intake of aspirin, alcohol, or certain foods typically exacerbates this burning discomfort, anti the discomfort may he relieved promptly by antacids or even by one or two swallows of food or water. Patients mav have accompanying Ivsplntp.M. n*piir)iitffii of oiiiliprsfril food. of weight lss The* symptoms of a hiatus hernia fend to be exacerbated by lying down, and all forms of acid-peptic disease may be worse in the early morning when acidic secretions arc not neutralized by food. Esophageal spasm, which may be induced by reflux of gastric acid into an esophagus in which the mucosa has been previously irritated, can cause a squeezing pain that may be indistinguishable from myocardial ischemia and that may even have a similar pattern of radiation. Pain resulting from gastric or duodenal ulcer (Chap. 252) is epigastric or substemal, usually commences about I to lih after meals, and is usu^ly relieved ill several minutes by antacids or milk. The discomfort caused by acute cholecystitis is more commonly described as an ache, which may be epigastric or substemal. It most commonly tends to occur an hour or so after meals and not in relation to exertion.

Emotional disorders are also commonly associated with chest pain. Usually, the discomfort is experienced as a sense of tightness, sometimes called aching," and occasionally it may be sufficiently s evere as to be designated a pain of considerable magnitude. Since the discomfort may be described as a tightness or constriction and is often localized at least in part beneath the sternum, it is not surprising that this type of discomfort is frequently confused with that of myocardial ischemia. Ordinarily, it lasts for a half hour or more, is unrelated to exertion, and with slow fluctuation of intensity. The association with fatigue or emotional strain is usually clear, although this may not be volunteered by the patient. Associated hyperventilation cun enuse innocent changes in the T waves and ST segments, which cun ho confused with comnnry artery disease.
APPROACH TO THE PATIENT WITH CHEST DISCOMFORT

A detailed and

meticulous history of the behavior of the pain is the cornerstone of the

evaluation. The location, radiation, quality, intensity, and duration of the episodes arc important. Even more so is the story of the aggravating and alleviating factors. A history of intense aggravation by breathing. coughing, or other respiratory niiivrmriils will usually point luwaii I tin* pk'iirn mul pericardium or

mediastinum as the site, although chcst wall pain is likewise affected by respiratory motion. Similarly, a pain that regularly appears on rapid walking, or with other exertion such as sexual activity, and vanishes a few minutes after stopping suggests the diagnosis of angina pectoris, although a similar story will occasionally be obtained from patients with skeletal disorders. While data from the history are of cardinal importance in the assessment of chcst discomfort, physicians should not be misled into overreliancc on any single feature. For example, acute myocardial infarction sometimes presents with pain that may be described as burning or even as sharp and may not be principally located in the substemal area.

CHAPTER 10 QUANTITATIVE ASPECTS OF CLINICAL REASONING

possibility of coronary disease, which may range from exceedingly unlikely in a young woman to (he 10 percenf range in a middle-aged man with many coronary risk factors. Diagnostic tests Although myocardial ischemia commonly is associated with electrocardiographic changes (Chap. 203), many patients have normal tracings between attacks, and some may even be normal during an episode of pain. However, depression of the ST segments, caused by myocardial ischemia, typically occurs during exertion and is accompanied by anginal discomfort: moreover, electrocardiographic evidence of myocardial ischemia may occuf at jpe'S't and with or without accompanying chest discomfort. The finding of flat or down-sloping ST-segmenl depressions of 0.1 mV or greater during an attack of pain substantially increases the likelihood that the pain is anginal in origin. Exercise electrocardiography^'will show ischcmic changes in about 50 to 80 percent of persons with symptomatic coronary disease but also in about 10 to 15 percent of patient^ who do not have coronary disease. The accuracy of ambufatory ischemia monitoring in the general population is less dear. Exercise thallium scintigraphy (Chap. 190) will demonstrate a perfusion defect in about 75 to 85 percent of patients with angina pcetoris and will be falsely positive in about 10 percent of patients who have chest discomfort from noncoronary causes.

43

A thorough physical examination can provide important clues to the cause of chest discomfort. Blood pressure should be checked in both arms if aortic dissection is being considered. Examination of the skin may reveal cyanosis, which suggests hypoxemia from cither diminished cardiac output or impaired respiratory function, or xanthel- asthma, which wouldsuggest hyperlipidemia and associated coronary disease. The finding of lymphadenopathy suggests a tumor. The examination of the chest wall should include both inspection and palpation to search for costochondritis and other musculoskeletal abnormalities. Lung examination may reveal a pleural rub, signs of pneumonic consolidation, or evidence of congestive heart failure. The physical examination may be totally normal in persons with severe myocardial ischemia, but it also may demonstrate abnormalities of vital signs, a third or fourth heart sound, or mitral regurgitation Frojji papillary muscle dysfunction. Aortic stenosis will be accompaniedJ?y its typical murmur (Chap. 201). The cardiac examination also should search for an increased pulmonic second sound that may indicate elevated pulmonary artery pressure, such as is found in pulmonary embolism, and the pericardial friction rub that strongly suggests pericarditis. A careful upper abdominal examination may be the first clue to peptic ulcer disease or cholecystitis. Critical information can often be obtained by attempts to produce or alleviate the pain, such as with nitroglycerin. Careful palpation of the chest wall, subacromial bursa, deltoid tendon, abdomen, and other structures may be very helpful if it reproduces the chcst discomfort. Shoulder and arm motion commonly reproduces pain related to these structures. However, the finding that such maneuvers can cause chest discomfort \loes not mean that such musculoskeletal diseases are the cause of the presenting complaint unless one can b. sure that the patients syndrome is reproduced precisely. Alternatively, the demonstration that a localized pain can be completely relieved by infiltration of a local anesthetic will be conclusive in convincing both the patient and the physician. Evaluation of the patient at the time of- a spontaneous episode, such as withnrn electrocardiogram during pain, is also extremely helpful.
APPLICATION OF THE PRINCIPLES OF CLINICAL REASONING The assessment of the probability of the various causes of chest pain requires the integration of multiple pieces of data, because no single clinical feature can be considered decisive. Each of the conditions that can cause chest discomfort can have varied (presentations, and the diagnostic tests upon which physicians often rely can also have falsepositive or false-negative results. Thus the principles of clinical reasoning (Chap. 10) should be applied to the evaluation of the patient with chest discomfort.

The evaluation of patients with suspected pulmonary embolism should usually focus on the documentation of deep venous thrombosis (Chap. 226) and the evaluation of pulmonary perfusion with a lung scintigram and/or pulmonary arteriography (Chap. 216). Aortic dissection is often suggested by the routine chcst radiograph, and the diagnosis may be established by echocardiography (especially transesophageal echocardiography), computed tomography, or magnetic resonance imaging. Aortography is the definitive test, but because__of^ its i^yasiveness, it is usually reserved for situations *ln which the suspicion of dissection is moderate or high and definitive anatomic documentation or localization is needed, often because of the need to consider a surgical repair. Esophageal or peptic ulcer diseases can often be diagnosed by an upper gastrointestinal roentgenogram. Esophageal manometry and measurement of lower esophageal sphincter pressure are useful in identifying esophageal spasm. 1'he Bernstein acid perfusion test, in which an attempt is made to reproduce die pain by infusing hydrochloric acid into the esophagus, can help establish acid reflux as the cause of pain (Chap. 251). Integration of clinical data and test results It is often useful to subdivide patients into those with an acute onset of a new or worsened chcst pain syndrome versus those with more chronic pain. Acute chcsl pain, with a duration of minutes to hours prior to the patient's presentation to a physician, could be caused by many of the entities described in this chapter and would be especially suspicious for acute myocardial infarction, aortic dissection, pulmonary embolism, biliary colic, or acute musculoskeletal trauma. In many situations, the patient

History and physical examination The information obtained from a careful incdical history and physical examination can be used to develop a differential diagnosis of the causes of chcst discomfort in an individual patient, to rank these diagnostic possibilities, ancT often to assign approximate percent probabilities to them. Although the various causes of chcst discomfort have typical characteristics, these characteristics must be interpreted in light of the prior probability that a person with a given age and sex and with a particular past mcdical history would have such a cause of chcst discomfort. For example, the possibility of angina pectoris as a cause of precordial or substemal discomfort must be seriously considered in a middle- aged man with coronary risk factors such as hypercholesterolemia and smoking, even if the description of the discomfort is not perfectly typical for angina pectoris. Conversely, wnen a 20-year-old woman describes the onset of new discomfort in a way that is seemingly classic for angina pectoris, such a diagnosis is relatively unlikely bccausc the prior probability of ischemic heart disease, given her age and sex, is so low. Although it is not always possible to assign numerical probabilities to the various causes of chcst discomfort in an individual patient, experienced clinicians either implicitly or explicitly assess the relative likelihoods of various potential explanations for any chest discomfort syndrome to help guide their future diagnostic evaluations and therapy. For example, a middle-aged or elderly man with typical characteristics for angina pectoris has about an 80 to 85 percent probability of having hemodynamically significant coronary artery disease. By comparison, the same man with a history of chcst discomfort that has some characteristics that arc typical for angina pectoris but other characteristics that arc atypical will have a probability of important coronary disease ranging from about 30 to 60 percent. Even persons with chest pain that is decidedly unlikely to represent coronary disease still have some finite

aortic aneurysm. This pain may persist over a period of several days before rupture and collapse occur. Abdominal wall Pain arising from the abdominal wall is usually constant unci aching. Movement, prolonged standing, and pressure accentuate the discomfort and muscle spasm. In the case of hematoma of the rectus sheath, now most frequently encountered in association with anticoagulant therapy, a mass may be present in the lower quadrants of the abdomen. Simultaneous involvement of muscles in other parts of the body usually serves to differentiate myositis of the abdominal wall from an intraabdominnl process which might cause pain in the same region.

CHAPTER 10 QUANTITATIVE ASPECTS OF CLINICAL REASONING

abnormalities, an underlying organic problem should be strongly suspected. Black widow spider bite^ produce intense pain and rigidity of the abdominal muscles and of the back, an area infrequently involved in disease of intraabdominal origin. NEUROGENIC CAUSES Causalgic pain may occur in diseases that injure nerves of sensory type. It has a burning character and is usually limited to the distribution of a given peripheral nerve. Normal stimuli such as touch or change in temperature may be transformed into this type of pain, which is also frequently present in a patient at rest. A helpful finding is the demonstration that cutaneous pain spots are now irregularly spaced, and this tnay be the only indication of an old nerve lesion underlying causalgic pain. Even though the pain may be precipitated by gentle palpation, rigidity of the abdominal muscles is absent, and the respirations are not disturbed. Distention of the abdomen is uncommon, and the pain has no relationship to the intake of food.

43

REFERRED PAIN IN ABDOMINAL DISEASES Pain referred to the abdomen from the thorax, spine, or genitalia may prove a vexing diagnostic problem, because diseases of the upper part of the abdominal cavity such as acute cholecystitis or perforated ulcer are frequently associated with intrathoracic complications. A most important, yet often forgotten dictum is that the possibility of intrathoracic disease must be considered in every patient with abdominal pain, especially if the pain is in the upper part of the abdomen. Systematic questioning and examination directed toward detecting the presence or absence of myocardial or pulmonary infarction, pneumonia, pericarditis, or esophageal disease (the intrathoracic diseases which most often masquerade as abdominal emergencies) will often provide sufficient clues to establish the proper diagnosis. Diaphragmatic pleuritis resulting from pneumonia or pulmonary infarction may cause pain in the right upper quadrant and' pain in the supraclavicular area, the latter radiation to be sharply distinguished from the referred subscapular pain caused by acute distention of the cxtrahepatic biliary tree. The ultimate decision as to the origin of abdominal pain may require deliberate and planned observation over a period of several hours, during which time repeated questioning and examination will provide the proper explanation. Referred pain of thoracic origin is often accompanied by splinting of the involved hemithorax with respiratory lag and decrease in excursion more marked than that seen in the presence of intraabdominal disease. In addition, apparent abdominal muscle spasm caused by referred pain will diminish during the inspiratory phase of respiration; whereas it is persistent throughout both respiratory phases if it is of abdominal origin. Palpation over the area of referred pain in the abdomen also docs not usually accentuate the pain and in many instances actually seems to relieve it. The frequent coexistence of thoracic and abdominal disease may be misleading and confusing, so differentiation may be difficult or impossible. For example, the patient with known biliary tract disease often has epigastric pain during myocardial infarction, or biliary colic may be referred to the prccor- dium or left shoulder in a patient who has suffered previously from angina pectoris. For an explanation of the radiation of pain to a previously diseased area, see Chap. 11. Referred pain from the spine, which usually involves compression or irritation of nerve roots, is characteristically intensified by certain motions such as. cough, sneeze, or strain and is associated with hyperesthesia over the involved dermatomes. Pain, referred to the abdomen from the testicles or seminal vesicles is generally accentuated by the .slightest pressure on either of these organs. The abdominal discomfort is of dull aching chaucter and is poorly localized. METABOLIC ABDOMINAL CRISES Win of metabolic origin may simulate almost any other type of intraabdomipul disease. Here several mcchanism* may|bc at work. In certain instances, such as hyperlipemia, the metabolic disease itself may be accompanied by an intraabdominal process such as pancreatitis, which can lead to unnecessary laparotomy unless recognized. C'l esterase deficiency sisMiciiiicil will) angiouctnohc edema is also often associated with episodes of severe abdominal pain. Whenever the cause of abdominal p:iin is ohsciin*. a mctabolic origin always must IK* considered. Abdominal pain is also the hallmark of familial Mediterranean fever (('Imp. 2H). ITic problem of differential diagnosis is often not readily resolved. The pain of porphyria and of lead colic usually is difficult to distinguish from that of intestinal obstruction, because severe hyperperistaJsis is a prominent feature of both. The pain of uremia or diabetes is nonspecific, and the pain and tenderness frequently shift in location and intensity. Diabetic acidosis may be precipitated by acute appendicitis or intestinal obstruction, so if prompt resolution of the abdominal pain docs not result from correction of thi metabolic

Pain arising from spinal nerves or roots comes and goes suddenly and is of a lancinating type (see Chap. 15). It may be caused by herpes zoster, impingement by arthritis, tumors, herniated nucleus pulposus, diabetes, or syphilis. Again, it is not associated with food intake, abdominal distention, or changes in respiration. Severe muscle spasm, as in the gastric crises of tabes dorsalis, is common but is either relievedjpr is not accentuated by abdominal palpation. The pain is made worse by movement of the spine and is usually confined to a few dermatome segments. Hyperesthesia is very common. Psychogenic pain conforms to none of the aforementioned patterns of disease. Here the mechanism is hard to define. The most common problem is the hysterical adolescent or young person who develops abdominal pain and who frequently loses an appendix or other organs because of it. Ovulation or some other natural event that causes brief mild abdominal discomfort may sometimes be experienced as an abdominal catastrophe. Psychogenic pain varies enormously in type and location but usually has no relation to meals. It is often at its onset markedly accentuated during the night. Nausea and vomiting are rarely observed, although occasionally the patient reports these symptoms. Spasm is seldom induced in the abdominal musculature and, if present, does not persist, especially if the attention of the patient can be distracted. Persistent localized tenderness is rare, and if found, the muscle spasm in the area is inconsistent and often absent. Restriction of the depth of respiration is the most common respiratory abnormality, but this is in the nature of a smothering or choking sensation and is part of an anxiety state. It occurs in the absence of thoracic splinting or change in the respiratory rate. APPROACH TO THE PATIENT WITH ABDOMINAL PAIN There are few abdominal conditions that require such urgent operative intervention that an orderly approach need be abandoned, no matter how ill the patient. Only those patients with exsanguinating hemorrhage must be rushed to the operating room immediately, but in such instances, only a few minutes arc required to assess the critical nature of the problem. Under these circumstances, all obstacles must be swept uside, adequate access for intravenous fluid replacement obtained, and the operation begun. Many patients of this type have died in the radiology department or the emergency room while f awaiting such unnecessary examinations us electrocardiograms or films of the abdomen. There are no contraindications to operation I 'lien, massive hemorrhage is present. Although exceedingly important, this situation fortunately is relatively rare. Nothin;* will supplant iiitfordeily, painstakingly detailed historyt which is far more valuable than any laboratory or roentgenologic

examination. This kind of history is laborious and time-consuming, making it not especially popular, even though a reasonably accurate diagnosis can be made on the basis of the history alone in the majority of eases. Recent studies of computer-aided diagnosis of abdominal pain indicate that this technique provides no advantage over clinical assessment alone. In cases of acute abdominal pain, a diagnosis is readily established in most instances, whereas success is not so frequently achieved in patients with chronic pain. Since the irritable bowel syndrome is one of the most common causes of abdominal pain, the possibility of this diagnosis must always be kept in mincer (see Chap. 256). The chronological sequence of events in the patients history is often more important than emphasis on the location of pain. If the examiner is sufficiently open-minded and unhurried, asks the proper questions, and listens, the patient will usually provide the diagnosis. Careful attention should be paid to the extraabdominal regions which may be responsible for abdominal pain. An accurate menstrual history in a female patient is essential. Narcotics or analgesics should be withheld until a definitive diagnosis or a definitive plan has been formulated, because these agents often make it more difficult to securc and to interpret the history and physical findings. In the examination, simple critical inspection of the patient, e.g.. of facies, position in bed, and respiratory activity, may provide valuable clues. The amount of information to be gleaned is directly proportional to the gentleness and thoroughness of the examined. Once a patient with peritoneal inflammation has been examined brusquely, accurate assessment by the next examiner becomes almost impossible. For example, eliciting rebound tenderness by sudden release of a deeply palpating hand in a patient with suspected peritonitis is cruel and unnecessary. The same information can be obtained by gentle percussion of the abdomen (rebound tenderness on a miniature scale), a maneuver which can be far more precise and localizing. Asking the patient to cough will elicit true rebound tenderness without the need for placing a hand on the abdomen. Furthermore, the forceful demonstration of rebound tenderness will startle and induce protective spasm in a nervous or worried patient in whom true rebound tenderness is not present. A palpable gallbladder will be missed if palpation is so brusque that voluntary muscle spasm becomes superimposed on involuntary muscular rigidity. As in history taking, there is no substitute for sufficient time spent in the examination. It is important to remember that abdominal signs may be minimal but nevertheless, if accompanied by consistent symptoms, may be exceptionally meaningful. Signs may be virtually or actually totally absent in cases of pelvic peritonitis, so careful pelvic and rectal examinations are mandatory in every patient with abdominal pain. The presence of tenderness on pelvic or rectal examination in the absence of other abdominal signs must lead the examiner to consider such important operative indications as perforated appendicitis, diverticulitis, twisted ovarian cyst, and many others. Much attention has been paid to the presence or absence of peristaltic sounds, their quality, and their frequency. Auscultation of the abdomen is probably one of the least rewarding aspects of the physical examination of a patient with abdominal pain. Severe catastrophes, such as strangulating small intestinal obstruction or perforated appendicitis, may occur in the presence of normal peristalsis. Conversely, when the proximal part of the intestine above an obstruction becomes markedly distended and edematous, peristaltic sounds may lose the characteristics of borborygmi and become weak or absent even when peritonitis is not present. It is usually the severe chemical peritonitis of sudden onset which is associated with the truly silent abdomen. Assessment of the patient's stale of hydration is important. The hematocrit and urinalysis permit an accurate estimate of the severity of dehydration so that adequate replacement can be carried out. Laboratory examinations may be of enormous value in assessment of the patient with abdominal pain, yet with but a few exceptions they rarely establish a diagnosis. Leukocytosis should never be the single deciding factor as to whether or not operation is indicated. A white blood cell count greater than 20,000/mm' may be observed with perforation of a viscus, but pancreatitis, acute cholecystitis, pelvic inflammatory disease, and intestinal infarction may be associated with marked leukocytosis. A normal white blood cell count is by no means rare in cases of perforation of abdominal viscera. The diagnosis of anemia may be more helpful than the whitcj^lood cell count, especially when combined with the history. The urinalysis is also of great value in indicating to some degree the state of hydration or to rule out severe renal disease, diabetes, or urinary infection. Determination of the blood urea nitrogen, blood sugar, and scrum bilirubin levels also may be helpful. The serum amylase determination is overrated. Since many diseases other than pancreatitis, e.g., perforated ulcer, strangulating intestinal obstruction, and acute cholecystitis, may be associated with very marked increase in the serum amylase, great care must be exercised

CHAPTER 10 QUANTITATIVE ASPECTS OF CLINICAL REASONING

in denying an operation to a patient solely on the basis of an elevated serum amylase level. The determination of the serum lipase may have a somewhat greater accuracy than the serum amylase.

43

Plain and upright or lateral decubitus roentgenograms of the abdomen may be of the greatest value. They are usually unnecessary in patients with acute appendicitis or strangulated external hernias. However, in cases of intestinal obstruction, perforated ulcer, and a variety of other conditions, films may be diagnostic. In rare instances, barium or water-soluble medium examination of the upper part of the gastrointestinal tract may demonstrate partial intestinal obstruction which may elude diagnosis by other means. If there is any question of obstruction of the colon, oral administration of barium sulfate should be avoided. On the other hand, barium enema is of inestimable value in cases of colonic obstruction and should be used with greater frequency where the possibility of perforation does not exist. Peritoneal lavage is a safe and effective diagnostic maneuver in patients with acute abdominal pain. It is of special^value in patients with blunt trauma to the abdomen, in whom evaluation of the abdomen may be difficult because of other multiple injuries to the spine, pelvis, or ribs and in whom blood in the peritoneal cavity produces only a very mild peritoneal reaction. In the absence of trauma, peritoneal lavage has been replaced by ultrasound and laparoscopy. Ultrasonography has proved to be useful in detecting an enlarged gallbladder or ^pancreas. the presence of gallstones, an enlarged ovary, or a tubal pregnancy. Laparoscopy is especially helpful in diagnosing pelvic conditions such as ovarian cysts, tubal pregnancies or salpingitis, and acute appendicitis. Radioisotopic scans (HIDA) may help differentiate acute cholecystitis from acute pancreatitis. A computed tomography (CT) scan may demonstrate an enlarged pancreas or a ruptured spleen, but it should be used only for specific questions such as these. Sometimes, even under the best of circumstances with all available auxiliary aids and with the greatest of clinical skill, a definitive diagnosis cannot be established at the time of the initial examination. Nevertheless, despite lack of a clear anatomic diagnosis, it may be abundantly clear to an experienced and thoughtful physician and surgeon that on clinical grounds alone operation is indicated. Should that decision be questionable, watchful waiting with repeated questioning and examination will often elucidate the true nature of the illness and indicate the proper course of action.

REFERENCES
D.wn.s All ct al: Ultrasonography m the acute abdomen. Ur J Surg 7S:1178. 1991 Li.i:. PWR: Tin* plain x-ray in the acute abdomen: A surgeon's evaluation. Hr J Surg ft3;7(3, pH Li-.i.k HI1": Abdominal and pelvic visceral rcccptors. Br Med Bull 33:163, 1977 Sil.EN W: 15th cd. London. Oxford Press. 1991

Sutton GC: How is I Lancet 2:905. 1989 Vaiman J 282:IH58, 19X1

liarly Diagnosis iif (he Acme AMimcn,
accurate pain. Br Med

Cope's

computcr-aidct# diagnosis? IB: Acute abdominal

CHAPTER 10 QUANTITATIVE ASPECTS OF CLINICAL REASONING

14 HEADACHE
NEIL H. RASKIN Few of us arc spared the experience of head pain during our lifetimes; indeed, severe, disabling headache is reported to occur at least annually by 40 percent of individuals worldwide. This incidence, occurs whether subjects live in large urban environments or in rural villages. The mechanism generating such "benign headaches may be activated by stress and anxiety, but emotional factors arc not necessary for thc#symptom to occur. The more severe the headache, the more likely it is to be associated with nausea and to be cxpcricnccd as a pulsing or pounding discomfort; photo- and phonophobia arc also more likely to be reported. Moreov er, there does no appear to be any utility in having headache; most sufferers report the contrary.

brain tumors.

43

normal aspect of living; it then follows that the mechanism for such a common

Since headache is a ubiquitous symptom, it may properly be regarded as a

Headaches that bear a relationship to certain biologic events or to physical environmental changes are essential data for triage of patients. The following exacerbating phenomena make the benign nature of the syndrome highly probable: provocation by red wine, sustained exertion, organic odors, hunger, lack of sleep#weather change, and menses. The association of diarrhea with attacks (Table 14-1) is pathognomonic of a benign disorder (migraine). The cessation oi, amelioration of headache during pregnancy, especially the second and third trimesters, is also pathognomonic of migraine. Patients with continuous benign headaches often observe a pain-free interlude of several minutes upon awakening before head pain commences. This phenomenon occurs with other centrally mediated pain syndromes, such as thalamic pain, but docs not occur among patients with somatic disease as the cause of pain. In attempting to elicit this information, patients commonly respond negatively to initial inquiries in the mistaken belief that because the relationships are not consistent, the validity of the observation is in question. Activation of the mechanism by red wine and hunger, for example, is always inconsistent, for reasons that are unclear. It is important to make this context clear to the patient or else valuable information may be lost, resulting in unnecessary neuroimaging. A history of amenorrhea or galactorrhea should lead one to question whether the polycystic ovary syndrome or a prolactin- secreting pituitary adenoma is the source of headache. Headache arising de novo in a patient with known malignancy suggests cither cerebral mctastascs or carcinomatous meningitis. When there is striking accentuation of pain with eye movement, a systemic infection and particularly meningitis should be seriously considered. Head pain appearing abruptly after bending, lifting, or coughing can be the clue to a posterior fossa mass or the Arnold Chiari malformation. Orthostatic headache arises after lumbar puncture and also occurs with subdural hematoma and benign intracranial hypertension. The

phenomenon is more likely to be ordinary than extraordinary. This ordinary headache-generating mechanism appears to be influenced by hereditary factors that may turn up the gain, resulting in susceptibility to more frequent or more severe head pain. The term migraine is used nowadays to refer to such a mechanism, in contrast to its prior usage referring to an aggregation of certain symptoms. Headache is usually a benign symptom and only occasionally is the manifestation of a serious illness, such as brain tumor or giant, ccll arteritis. Thq. first issue to resolve in confronting the patient who complains of headache is to make the distinction between benign and more ominous causes. * GENERAL CONSIDERATIONS The quality, location, duration, and time course of the headache and the conditions that produce, exacerbate, or relieve it should be carefully reviewed with the patient. Ascertaining the quality of cephalic pain is occasionally helpful. Most headaches are dull, deeply located, and of aching character. Superimposed on such nondescript pain may be other pain elements tfiat have greater diagnostic value. It is useful to clarify to the patient that it is of interest to learn about all the pain elements that have been experienced regardless of their frequency or intensity. A throbbing # quality and tight muscles about the head, neck, and shoulder girdle arc common nonspecific accompaniments of headache, suggesting that intra- and cxtracranial arteries and skeletal muscle surrounding the.head and neck are activated by a generic head pain-generating, mechanism. It was formerly believed that tight hat-band headachep indicated anxiety or depression,vbut investigations have not supported this view. Jabbing, brief, sharp cephalic pain, often occulting multifocally (ice picklike pain), is the signature of a benign disorder. Pain inii'iisity seldom has diagnostic valuein the head or in any other somatic location. Prom tlu*' therapeutic perspective, it is, of course, the single as|>ecl of pain that is most important. Physicians should be cautious#ibout assessing pain intensity by visually inspecting a patient. People respond to pain in a variety of ways that range lrom overt histrionic behavior to stoicism. Inquiries as to how pain disturbs day-to-day function may extract more useful information. Response to placebo nieilicaiion or procedures produces no useful injonnation either diagnostic or therapeutic. It simply identifies a placebo responder, about 30 percent of the population. There is no evidence that placebo responders have lower pain levels than nonresponders or do not really have pain. Patients entering emergency department with the most severe headache of their lives usually have migraine. Meningitis, siihaiacltnoitl In'iuoiiliafc. and clu.stci headache also produce intense cranial pain. Contrary to common belief, the headache produced by a brain tumor is not usually particularly or distinctively ftcvcrc. Oat a rci'ardiii}* liHiiiimi of headachc may be informative. If the Mturcc is an cxtracranial structure, as in giant cell arteritis, the om with tin* mi of fairly precise Inflammation

TABLE 14-1 Symptoms accompanying severe migraine attacks in 500 patients

Percentage affected
87

56 16 82 36

10
26

33
65 72 33 18 4 4
11: Nil Kn tkm. I/re*

10
hnbr. M t. Now Ytk. (Ii'inlull l.itingMone, IWX.

yc *tscll is seldom the cause of aeulc orbital pain if die sclerae are while and not injcctcd; a red eye is die sign ol ophthalmic disease. Similarly, acute sinusitus nearly always declares1 itself through a dark green, purulent nasal exudate.
Symptom
Nausea Vomiting Diarrhea Photophobia Visual disturbances Ponilication s|>ccU'a Photopsia Paivsthi'sias Nvalp lomlciness Lighihcailetlness Vertigo Alteration of consciousness Sc i /.ti iv Sytieiipc

of an extrvemnfaf artery causes pain and cx tcmtiva iMrfmM to the site of the
vessel. Lesions of paranasal sinuses, teeth, eves, and upper cervical vertebrae induce less sharply localized pain, but one that is still referred in a regional distribution that is quite constant. Intracranial lesions in the posterior fossa cause pain that is usually occipitonuchal, and supratentorial lesions most often induce fronto- tcmporal pain, __ '

Duration and time-intensity curves of headaches are particularly useful. A ruptured aneurysm results in head pain that peaks in an instant, thunderclaplike; much less often, unmptured aneurysms may signal their presence in the same way. Ouster headache attacks reach their peak over 3 to 5 min, remain at maximal levels for about 45 * min, and then taper off. Migraine attacks build up over hours, are maintained for several hours to days, and are characteristically relieved by sleep. Sleep disruption is characteristic of headaches produced by

The analysis of facial pain requires a disparate approach. Trigeminal and glossopharyngeal neuralgia are common causes of facial pain, especially the former. Neuralgias" arc painful disorders characterized by paroxysmal, fleeting, often electric shocklike episodes (hat are caused by dcmyelinativc lesions of nerves (the trigeminal or glossopharyngeal nerves in cranial neuralgias) that result in the activation of a CNS pain..mal Male generating mechanism Certain maneuvers characteristically^ trigger paroxysms of pain. However, the most common cause of facial pain by far is dental; provocation by hot, cold, or sweet foods is typical. The application of a cold stimulus will repeatedly induce dental pain whereas in neuralgic disorders a refractory period usually occurs after the initial response so that pain cannot be repeatedly induced. The presence of refractory periods can nearly always be elicited in the history so that patients need not be put through a painful experience. #

Mealtimes offer the physician an opportunity to gain needed insight into the mechanism of a patient's facial pain. Is it the chewing, the swallowing, or the taste of the food that elicits pain? Chewing points toward trigeminal neuralgia, temporomandibularjoint dysfunction i or giant cell arteritis (jaw claudication), whereas swallowing and taste provocation points toward glossopharyngeal neuralgia, fcain upon swallowing is common among patients with/carotidynia) (facial migraine, sec below) because the inflamed, tender carotid artery abuts the esophagus during deglutition. ;

CHAPTER 10 QUANTITATIVE ASPECTS OF CLINICAL REASONING

Many patients with the complaint oT facial pain do not describe stereotypic syndromes, in parallel with most painful conditions; such patients have sometimes had their syndromes categorized as atypical facial pain as if this were a well-defined clinical entity. There is only scant evidence that nondescript facial pain is caused by emotional distress, as has sometimes been alleged. Vague, poorly localized,# continuous facial pain is characteristic of the condition that may result from nasopharyngeal carcinoma and other somatic diseases; a burning painful element often supervenes as dcaffcrcntation occurs and evidence of cranial neuropathy appears. Occasionally, the cause of a pain problem cannot be resolved promptly, necessitating periodic follow-up until further clues appear (and they usually do). Facial pain of unknown cause'' appears to be a more reasonable tentative diagnosis than atypical facial pain.'' PAIN-SENSITIVE STRUCTURES OF THE HEAD The most common type of pain is that resulting from activation of peripheral nociceptors in the presence of a normally functioning nervous system, as in the pain resulting from scalded skin or appendicitis. Another type of pain is the result of injury or activation of the peripheral or central nervous system. Headache, formerly believed to originate peripherally, may originate from either mechanism. Headache may arise from dysfunction, displacement, or encroachment upon pain- sensitive cranial structures/The following are sensitive to mechanical Simulation: the scato and^aponeurotica, middjejmeningeal artery, dural sinuses, falx cerebri, and the proximal segments of the large pial arteries/The ventricular ependyma, choroid plexus, pial veins, ancTmuch o f m e brain parenchyma jire pain^lnsensitivcjOn the other hand, electric^'stimulation near midbrain dorsal raphe cells has resulted in migrainelikc headaches. Thus most of the brain is insensitive to electrode probing, but a particular midbrain site is nevertheless a putative locus for hcadachc generation Sensory stimuli from the head arc conveyed to the central nervous system via the trigeminal nerves for structures above the tentorium in the anterior and middle fossae of the skull and via the first three cervical nerves for those in the posterior fossa and infradural structures. The ninth and lentil cranial nerves supply part of the posterior fossa and refer pain to the ear and throat. Hcadachc can occur as the result of (I) distention, traction, or dilation of intracranial or extracrania! arteries; (2) traction or displacement of 1 arge J n t

tenderness. Milder headaches tend to be nondescripttight, bandlike discomfort often involving the entire .head the profile of 'tension headache. These differing clinical profiles of headaches that are not caused by an intracranial structural anomaly g or systemic disease probably represent different points on a continuum rather than disparate clinical entities. Whether a single common, mechanism underlies these varying headache profiles is not entirely clear and remains to be investigated further. A working definition of migraine offered here is benign recurring headache and/or neurologic dysfunction usually attended by pain-free interludes and almost always # provoked by stereotyped stimuli. It is by far more common in women; there is a hereditary predisposition toward attacks; and the cranial drculatory phenomena that attend attacks appear to be secondary to a primary CNS disorder.

43

Clinical subtypes The designation classic migraine (migraine with aura) denotes the syndrome of headache associated with characteristic premonitory sensory, motor, or visual symptoms; com/hon migraine (migraine without,aura) denotes one in which them is no focal neurologic disturbance preceding the occurrence of headache. However, the latter is by far the most frequent clinical problem, and focul neurologic disturbances arc more common during headache attacks than as prodromal symptoms, Focal neurologic disturbances without hcadachc or vomiting have come to.be known as migraine equivalents or accompaniments and appear to occur more commonly in patients between the agi;s ol 40 and 70 year*. Hie term complicated migraine has generally been used to describe.migraine with dramatic local neurologic features, athus overlapping with classic migraine: it has also been used to connote a persisting neurologic dcficit that is a residuum of a migraine attack. the prevalence of migraine among the elderly is substantial, considerably higher than giant cell arteritis. Cough headache One of the male-dominated (4:1) syndromes, it is characterized by transient, severe head pain upon coughing, bending, lifting, sneezing, or stooping. Head pain persists for seconds to a few minutes. Many patients date the origins of the syndrome to a lower respiratory infection accompanied by severe coughing or to strenuous weight-lifting programs. Headache is usually diffuse but is lateralized in about one-third of patients. The incidence of serious intracranial structural anomalies causing thisj condition is about 25 percent; the ArnokJ-Chiari malformation i,s a common cause. Magnetic resonance imaging is indicated for most of these patients. The benign disorder may persist for a few years; it is inexplicably and remarkably ameliorated by indomcthacin at doses ranging from 50 to 200 mg daily. Many patients with migraine note that attacks of headache may be provoked by sustained physical exertion, such as during the third mile of a 5-mile run. Such headaches build up over hours, distinctly different from the cough headache syndrome. The term effort migraine has been used for this syndrome to avoid the ambiguous term exertional headache. Coital headache Another male-dominated (4:1) syndrome, attacks occur periorgasmically, are very abrupt in onset, and subside, in a few minutes if coitus is interrupted. These are nearly always benign events and usually occur sporadically; if they persist for hours or are accompanied by vomiting, subarachnoid hemorrhage must be excluded through a CSF examination and CT scanning. Brain tumor headache About 30 percent of patients with brain tumors consider headache to be their chiefi complaint. The head pain syndrome is nondescript; a deep, dull ncniug quality, of moderate intensity, occurs intermittently, is worsened by exertion or change in position, and is associated with nausea and vomiting. This pattern of symptoms results from migraine far more often than from brain tumor. Headache disturbs sleep in about 10 percent of patients. Vomiting that precedes the appearance of headache by weeks is highly characteristic of posterior fossa brain tumors. Headache caused by systemic illness"* There is hardly any illness that is never manifested by headache; however, some illnesses arc characteristically associated with headache. These include infectious mononucleosis, systemic lupus erythematosus, chronic pulmonary failure with hypercapnia (early morning headaches), Hashimotos thyroiditis, glucocorticoid withdrawal, oral contraceptives, ovulation- promoting medications, inflammatory bowel disease, many of the HIV-associatecUllnesses, and the acute blood pressure elevations that occur in pheochromocytoma and in malignant hypertension. The last two examples are the exceptions to the generalization that hypertension per sc is a very uncommon cause of headache; diastolic pressures of at least 120 mmHg are requisite for hypertension to cause headache. APPROACH TO THE PATIENT WITH HEADACHE Entirely different diagnostic possibilities are raised by a patient who presents with the first severe headache ever and a patient who has had recurrent headache over many years. In the first instance, the probability of finding a potentially serious cause is considerably greater than in the second; some of the causes that should be considered include meningitis, subarachnoid hemorrhage, epidural or subdural hematoma, glaucoma, and purulent sinusitis. In general, acute, severe headache with stiff neck and fever means meningitis and without fever means subarachnoid hemorrhage; when confronted with such a patient, lumbar

racrani a|ve i n s o r t h c i r d u r a I envelope; (3) compression, traction, or inflammation of cranial and spinal nerves; M) Spasm, inflammation, and trauma lo cranial and cervical inuuln; (5) meningeal irritation and raised intracranial pressure: and (6) perturbation of intracerebraf .serotonergic projections. By and large, intracranial masses cause headache when they deform, displace, or exeil If action on vessels, (filial sliut fines, oi cranial nerves al (lie base ol the brain* this often happens long before intracranial pressure rises. Suefi mechanical displacement nicelianisms do mil explain (he headaches resulting from cerebral ischemia, or from benign intracranial hypertension after the pressure is reduccd. or the hcadacbes that are so common in febrile illnesses and systemic lupus erythematosus. Perturbation of intracercbral serotonergic projections has. been posited as a possible mechanism for these phenomena.

PRINCIPAL CLINICAL VARIETIES OF HEADACHE

Normally there is little difficulty in diagnosing the headache of glaucoma, purulent sinusitis, bacterial meningitis, and brain tumor because of the clues provided by the associated symptoms and signs. Hcadachc alone is nondescript. It is wlien headache is chronic, recurrent, and unattended by other important signs of disease that the physician faces a challenging but ultimately gratifying medicaJ problem. The headache syndromes described below should be considered (see Table 14-2). MIGRAINE; The term migraine stems from Galen's usage of liemierania loyicscribc a periodic disorder comprising paroxysmal blinding heitiicramal pain, vomiime. photophobia* recurrence at regular intervals, and relief by dark surroundings and sleep. Hemicra- nia was later corrupted into low Latin as hemigranea and migranea\ eventually the French translation, migraine, gained acceptance in the eighteenth century and has prevailed ever since. The passage of time has proved this to be a misleading designation for a condition manifested by latcralizcd head pain in less than 60 percent of those affected. Furthermore, undue emphasis on the dramatic features of migraine has often led to the illogical conclusion that periodic hcadachc#lacking such features is not migrainous in mechanism. It has become clear that severe headache attacks, regardless of cause, are more likely to be described as throbbing an<^ associated with vomiting and scalp

puncture is mandatory. Acute persistent headache and fever is often the manifestation of an acute systemic viral infection; if the neck is supple in such a patient, lumbar puncture may be deferred. There is always the possibility of a first attack of migraine, but fever would be a rare associated Icaturc.

CHAPTER 10 QUANTITATIVE ASPECTS OF CLINICAL REASONING

ANATOMY AND PHYSIOLOGY OF THE LOWER PART OF THE BACK

43

When the signature of migraine has not clarified the cause of recurring headache, one should consider the investigation of cardiovascular and renal status by blood pressure and urine examination; eyes by fundoscopy, intraocular pressure measurement and refraction; the cranial arteries by palpation; the cervical spine by the
effect oJ' passive movement of the bead and imaging; and the

neurologic and psychological evaluation.

nervous

system by

The adolescent with chronic daily frontal or holoccphalic headache represents a special type of problem. Extensive diagnostic batteries are most often unrevealing, including the psychiatric assessment. Fortunately, the headaches tend to sfop after a few years so that structured analgesic support can enable these teenagers to move through secondary school and enter college. By the time they reach the late teens, the cycle has usually ended. The relationship of head pain to depression is not straightforward. Many patients in chronic daily pain cycles become depressed, a not unreasonable sequence of events; moreover, there is a greater-than- chancc coincidence of migraine with both bipolar (manic depressive) and unipolar depressive disorders. Studies of large populations of depressed patients do not reveal headache prevalence rates that are different from the general population. The physician should be cautious about assigning depression as the cause of recurring headache; drugs with antidepressant actions are effective in migraine
also.

Finally, a note on recurring headache that may be pain-driven. Temporomandibular, joint dysfunction is an example; in general, it produces preauricular pain that is associated with chewing food. The pain may radiate to the head but is not easily confused with headache per se. On the other hand, headache-prone patients may observe that headaches are more frequent and severe in the presence of a painful temporomandibular joint problem. Similarly, headache disorders may be activated by the pain attending otological or endodontic surgical procedures. Treatment of the headache problem is largely ineffectual until the cause of the primary pain problem is dealt with. Thus pain about the head as the result of somatic disease or trauma may reawaken an otherwise quiescent migrainous mechanism.

Bickekstakf Blau Boles Couch Dechant Clissold Goadshy Edvinsson Goadsiiy Gundlacii Hakdebo Hughes Jensen Johns Lance JW:
ER: Basilar artery migraine. Lancet 1:15, 1961 JN: Migraine: Theories of pathogenosis. Lancet: 339:1202, 1992. DB: Visual field effects of classical migraine. Brain Cogn 21:181. 1993 JR: Headache to worry about. Med Clin North Am 77:141. 1993 KL. SP: Sumatriptan. Drugs 43:776, 1992 PJ, L: The trigeminovascular system and migraineStudies characterizing ccrchrovascular and neuropeptide changes seen in humans and cats. Ann Neurol 33:48, 1993 PJ, AL: Localization of 3 H-dihydroergoiamine-binding sites in the cm ccntral nervous system: Rclevance tu migraine. Ann Neurol 29:91, 1991 JE: Subcutaneous sumatriptan in cluster headache A time study of the eflect on pain and autonomic symptoms. Headache 33:18, 1993 RL: Identification and treatment of cerebral aneurysms after sentinel headache. Neurology 42:1118, 1992 R et al: Muscle tenderness and pressure pain thresholds in headacheA population study. Pain 52:193 1993 D: Benign sexual headache within a family. Arch Neurol 43:1158. 1986 Mechanism and Management of Headache, 5th ed. London. Butterworth Scientific, 1993 ----- : Headaches related to sexual activity. J Neurol Neurosurg Psychiatry 39:1226, 1976 r/ cl al: Pain mechanisms underlying vascular hcadachcs. Rev Neurol (Paris) MS: 181, 1989 OLESEN J: Cerebral and extracranial circulatory disturbances in migraine: Pathophysiological implications. Cerebrova.se Brain Metab Rev 3:1, 1991 RANDO TA, FISHMAN RA: Spontaneous intracranial hypotension. Neurology 42:481. 1992 RASKIN NH: Pharmacology of migraine. Prog Drug RES 34:209, 1990 ------- : Lumbar puncture headache:. A review. Headache 30:197, 1990 ----- : Headache, 2d ed. New York, Churchill Livingstone, 1988 RASMUSSEN BK, Oli-sen J: Symptomatic and nonsymptomaiic headaches in a general population. Neurology 42:1225, 1992 Sjaastad O: Cluster Headache Syndrome. London, Saunders. 1992 Si AND PI- et al: Incidence of migraine headache: A population-based study in Otmsled (Anility. Minnesota. Neurology 42:1657, 1992. Vinken PJ, Bkuyn GW, Klawans HI. (eds): Handbook of Clinical Neurology, vol 48: Iti'iidarlir Amsterdam, I.Kevin Science, 1980

REFERENCES

The bony spine is anatomically divisible into two parts. The anterior part consists of a scries of cylindrical vertebral bodies connected to one another by the intervertebral disks and held tightly together by the anterior and posterior longitudinal ligaments. The posterior part consists of more delicate elements that extend from the vertebral body as pedicles and broaden posteriorly to form laminae, which together with ligamentous structures form the vertebral canal. The posterior elements are joined to adjacent vertebrae by two small facetal synovial joints which allow a modest degree of motion between any two segments but in aggregate produce a rather extensive range (Fig. 15-1). Stout transverse and spinous bony processes project laterally and posteriorly and serve as the attachments of nfuscles which move, support, and protect the vertebral column. The stability of the spine depends on. two types of support: that provided by the bony articulations (principally by the diskal joints and the synovial articulations of the posterior elements) and a second type provided by the ligamentous (passive) and muscular (active) supporting structures. The ligamentous structures arc quite strong, but bccausc neither they nor the vertebral body-disk complexes have sufficient integral strength to resist the enormous forces acting on the column during even simple movements, voluntary and reflex contractions of the sacrospinalis, abdominal, gluteal, psoas, and( hamstring muscles afford much of the stability. The vertebral and paravertebral structures are innervated by branches of the segmental spinal nerves that exit the neural foramina at each spinal level. The sinuvertebral nerve, which is considered the major sensory nerve supply to the structures of the lumbar spine, arises from the spinal nerve prior to its division into an anterior and posterior ramus. The sinu vertebral nerve reenters (he spinal canal through the intervertebral foramen to provide sensory innervation to the posterior longitudinal ligament, external portions of the posterior annulus fibrosus, anterior dura, dura of the nerve root sleeve, and epidural veins, all within the spinal canal. The other major nerve supply to the spinal and paraspinal structures arises from the posterior primary ramus. The posterior primary ramus of the spinal nerve further divides into medial and lateral branches. Together these nerves supply the posterior parts of the spine, including the facet joints, as well as the paraspinal muscles and fascia. In addition, the upper three lumbar spinal nerves provide cutaneous sensation to the skin of the low back.
FIGURE 15-1 Left: Superior view of a stripped lumbar vertebra. Right: Lateral view of two articulated lumbar vertebrae. B ~ body; SC = spinal canal; IVF intervertebral foramen; IF = inferior articular facet; SF = superior articular facet; P = pedicle; TP = transverse process; SP - spinous process; L = lamina. (Adapted from DD Levine, in Arthritis and Allied Conditions: A Textbook of Rheumatology, 10th ed, DJ McCarty (ed). Philadelphia, Isa Hrhigcr, /9H5.)

Moskowi MA

I In' parts of |hi' back that |msscss the greatest frivilom of movement, and hence are most frequently subject to injury, arc the lumbar and cervical regions. In addition to (he voluntary motion* required lor bending, twisting, and other movements, many actions ol I lie spine are reflex in nature and are the basis of posture, GENERAL CLINICAL CONSIDERATIONS TYPES OF LOW BACK PAIN Four types of pain may be differentiated; local, referred, radicular, and that arising from secondary (protective) muscular spasm. Local pain is caused by any pathologic process that impinges on or irritates sensory endings. Involvement of structures that contain no sensory endings is painless. The central, medullary portion of the vertebral body may be destroyed by tumor, for example, without evocation of pain, whereas cortical fractures or tears and distortions of the periosteum, synovial membranes, muscles, annulus iibrosus. and ligaments arc often exquisitely painful. The latter structures arc

15

BACK AND NECK PAIN

HENRY J. MANKIN / LAWRENCE F. BORGES

CHAPTER 10 QUANTITATIVE ASPECTS OF CLINICAL REASONING

innervated by afferent fibers of the posterior primary rami and ihc sinu vertebral nerve. Although painful .states are often accompanied by swelling of the affected tissues, this may not be apparent if a deep structure of the back is (he site ol disease. Local pain is often described as steady but may be intermittent, varying considerably with position oi^activity^JEhc pain may hr sharp ordull and although often diffuse is always felt in or near the affected part of the spine. Reflex splinting of the spine segments by paravertebral muscles is frequently noted and may produce deformity or postural abnormality. Certain movements or postures that alter (he position of the injured tissues aggravate the pain. Firm pressure or percussion on superficial structures in the region involved usually evokes tenderness, which is of aid in identifying the site of the abnormality. Referred pain is of two types: that projected from the spine into regions lying within the area of the lumbar and upper sacral dermatomes and that projected from the pelvic and abdominal viscera to the spine. Pain due to diseases of the upper pari of Ihc lumbar spine is usually referred to th<f anterior aspects of the thighs and legs; (hat from the lower lumbar and sacral segments js referred to the gluteal regions, posterior thighs, calves, and sometimes feet. Pain of this type, although of deep, aching quality and rather diffuse, tends at times to be superficially projected. In general, the referred pain parallels in intensity the local pain in the back. In other words, maneuvers that alter local pain have a similar effect on referred pain, though not with such precision and immediacy as in radicular, or root," pain. Referred pain may be confused with pain from visceral disease, but the latter is usually described as deep and tends to radiate from the abdomen through to the back. Also, visccral pain is usually unaffcctcd by movement of the spine, does not improve with recumbency, and may be modified by the activity of the involved viscus. An important exception to this is pain caused by an aortic aneurysm. A slowly enlarging aortic aneurysm may erode the anterolateral spine and produce discomfort that changes with movement or recumbency. Radicular, or root, pain has sonic of the characteristics of referred pain but differs in its greater intensity, distal radiation, circumscription to the territory of a root, and the factors which excite il (Table 15-1). The mechanisms arc principally disttMliou. stretching, irritation, and compression of a spinal root, most often central to the intervertebral foramen. In addition, it has been suggested that in patients with spinal stenosis the lumbar claudication" pattern may be due to a relative ischemia associated with compression. Although the pain itself is often dull or aching, various maneuvers which increase the irritation of the root or stretch it may greatly intensify the pain, eliciting a lancinating quality. Nearly always the radiation of pain is from a central position near the spine to some part of the lower extremity. Cough, sneeze, and strain are characteristic evocative maneuvers, but since they njay also jar or move the spine, they may aggravate local pain as well/Forward bending with the knees extended

43