The vast majority of headaches are of the tension, migraine, and cluster types, which are classified as primary h eadaches and are discussed in the previous chapter. Unfortunately, many patients develop refractory headaches, which usually consist of one or more primary headache disorders complicated by analgesic medication overuse, poor coping patterns, or failure to identify triggers. In such cases, an interdisciplinary management approach is needed.

Of particular concern to patients and clinicians are the secondary headaches , also known as organic headaches, accounting for fewer than 10% of all recorded headaches [1]. By definition, they are symptomatic of underlying disease, structural pathology, or pain-inducing processes different from those traditionally ascribed to the primary headaches. Organic headaches may be secondary to elevated cerebrospinal fluid (CSF) pressure, known as benign intracranial hypertension or pseudotumor cerebri; to bleeding from congenital aneurysms or arteriovenous malformations (AVMs); to ischemic or hemorrhagic stroke; as well as to pain caused by mass lesions or mass effect, such as tumors, hematomas, AVMs, and trauma, or infectious processes such as meningitis, encephalitis, and cerebral abscesses. The clinical importance of organic headaches despite their relatively low prevalence compared with primary headaches illustrates important principles of diagnosis and treatment useful to the clinician prior to obtaining specialist consultation.

The topic is complex, with potentially far-reaching consequences if errors in the workup are made. The important point for the evaluating clinician is to know when and what to look for when suspecting an underlying cause of headache, and how to evaluate it within the financial limits of today’s medical environment or within the constraints of the patient’s medical condition. This requires knowledge of headache presentation, the limits of clinical dogma, awareness of new or less commonly used tests, and a familiarity with imaging and other diagnostic studies, not to mention a compassionate bedside manner.

This chapter focuses on the clinical signs, symptoms, and diagnostic workup of selected categories of secondary headaches. Cerebral tumors, stroke, subarachnoid hemorrhage, vascular anomalies, spinal headache (i.e., spontaneous CSF leaks or those caused by lumbar puncture or epidural misplacement), and infection with the human immunodeficiency virus (HIV) are some of the vehicles used to discuss common clinical scenarios and approaches to decision making. The information that follows is based on both the published literature and our clinical experience as neurologists and headache and pain specialists.

Special emphasis is placed on the importance of maintaining patient comfort, dignity, and self-esteem. This approach is crucial to the success of therapy, even if success is not always defined as “cure”, and is especially important in cases of terminal disease or those in which the primary diagnosis or headache symptom is disrupting patient and family dynamics.

Differentiating secondary from primary headaches can be difficult. The quality of pain may be indistinguishable from that of migraine, tension-type, or other primary headaches. In such cases, the Inter National Headache Society (IHS) states that the temporal relationship between the headache and underlying pathology should be the deciding factor [2]. Preexisting headaches aggravated by an organic process are still considered primary. If the onset of headache occurs in close proximity to the underlying structural problem, it is considered secondary. Sometimes the question is merely academic or impossible to answer. For example, how does one classify long-standing, stereotypic, but side-locked migraines when a magnetic resonance imaging (MRI) scan of the head, obtained to evaluate the cause of new-onset seizures, reveals the presence of an AVM ipsilateral to the headache, and in a position to cause pain? Are the headaches then primary or secondary? Perhaps the more important question concerns the risk of hemorrhage and neurologic deficits if the malformed blood vessels are left alone, removed, or otherwise treated.

Patients (and parents) often seek consultation for their headaches in hope of reassurance that they (or their children) do not have an underlying disease of which pain is but one symptom. They also want to understand why they are experiencing headaches. This need to understand is of fundamental importance to many patients and their families, a point often missed by clinicians. In our experience, also supported by literature, patients’ need to understand their disease may surpass their need for reassurance that pain medicine will be made available or even that the pain can be relieved [3]. This point is surprising only if one assumes that most patients do not think much about the details of their condition or care. Failure to appreciate this concept may lead to poor communication, mutual loss of respect, and frustration on both sides, which can certainly wind up the level of pain or the frequency of headaches, unnecessarily increasing the patient’s suffering.

Pain and headache may either impact or be impacted by patient coping skills, expectation of outcome, or feelings of helplessness and hopelessness [4]. Clinicians need to incorporate the psychology of health and disease in their approach and not merely focus on the more tangible medical signs and symptoms. Patients are people, not diseases; and their fear of brain tumors, for example, often goes unstated. It is beneficial to broach the subject regardless of whether or not the pattern of headache raises a concern in the clinician’s mind. It is surprising how frequently patients breathe a sigh of relief when they find out why their headache is likely to be caused by a tumor. Even when a tumor is diagnosed, the patients’ anxiety and fear of the unknown can be lowered when they are given an understanding of the mechanism of pain appropriate to their level of interest, as well as knowledge of what to expect over time. This approach minimizes the likelihood of depression caused by the feeling of hopelessness and helplessness that accompanies escalating and misunderstood refractory or frequently recurrent pain.

Diagnostic headache evaluations can be a double-edged sword, especially with regard to ethical considerations. Treatment of structural pathology may prove preventative of serious problems such as stroke, seizures, or even death, and may minimize or eliminate the headaches. However, many identifiable structural anomalies may or may not be amenable to, or even appropriately subjected to, invasive intervention. Importantly, there may be no relationship between the pain and the anomaly found. Risk versus benefit must therefore be carefully considered before embarking on the “latest and the greatest” diagnostics. Tests alone may carry physical, emotional, and financial risk, leading to significant ramifications for the patient as well as for the family. Contrary to the opinion of many clinicians, patient refusal to consider invasive treatment is not, per se, reason for psychiatric consultation. Such refusal certainly merits a gentle, affirming, and understanding discussion though, perhaps over several sessions, in order for the patient to process the information at his or her own speed. Informed consent or refusal is not always accomplished by a 5-min distillation of the medical “facts” and the dual signing of a piece of paper.

Occasionally, stereotypic-sounding migraine headaches with aura, especially if side- locked (always starting on the same side or in the same place), may turn out to be secondary to underlying structural pathology, sometimes with an associated risk of bleeding and stroke. A good example is an AVM, in which veins connect directly to arteries without the usual intervening arterioles [5, 6]. Finding an AVM in a family member with migraine may prompt evaluations in relatives known to experience similar headaches, but who have not yet undergone a formal workup. Should all related family members with headache be similarly evaluated? What about those without headache? Whether or not such evaluations are warranted is a matter of controversy and depends on the pathology in question, the philosophy of the clinician, and the individual patient’s resources and wishes. In general, focal seizures , vascular bruits, equivocal computed tomographic (CT) scans, and episodes of hemorrhage increase the likelihood that studies will reveal a clinically important headache-related lesion.

Do AVMs cause headache? Some researchers consider them more likely than aneurysms to cause migraine-type symptoms. Even here, some investigators postulate AVMs and migraine to be but co-morbid conditions, with the main contribution of the AVM being that of potential cerebral ischemia or severely diminished blood flow leading to temporary or permanent neurologic sequelae. This process may affect the nature of the aura but is unlikely to be causative of pain, the aura being an independent process added to by the embarrassment of blood flow caused by the AVM. In animal models, and to some extent in humans, a slowly spreading electrical depression of cortical neuronal function appears to correlate with the migraine aura more so than simple blood flow changes. Although a brief leading wave of small vessel hyperperfusion followed by a more prolonged state of hypo-perfusion or oligemia may be associated with cortical neuronal depression, actual ischemia (more severe) is not usually observed during the aura [7]. The pathophysiology of migraine remains controversial and is discussed in more detail in the previous chapter.

Finding an abnormality such as an AVM may have unexpected social and behavioral ramifications . For example, it may place a “red flag” in the person’s medical record that could interfere with his or her ability to procure a change in, or an upgrade of, a health insurance plan. Furthermore, it could lead to patient hypervigilance over somatic sensations, causing anxiety and, therefore, more frequent headaches. The situation is analogous to MRI-discovered disk bulges in those with spinal pain. Such findings are common but may bear no causal relationship to the back pain. Patients may have trouble understanding the logic of the clinician’s recommendation that no invasive treatment is advised, so careful explanation is usually necessary. This takes the time that many clinicians feel they do not have. However, not taking the time to fully educate the patient may lead to more clinic phone calls and worsening of headaches due to stress. The patient may eventually find someone willing to operate. Paradoxically, for the “worried-well” patients, ordering a head scan, even if likely to be unremarkable, may actually prove cost-effective because of the value of reassurance. But the clinician should be aware that an unrelated anomaly might be found, avoiding the problematic cascade of events mentioned earlier. What, then, is the appropriate course of action when a structural lesion is found? Weighing the relative pros and cons of neurosurgical intervention in a headache-prone, but otherwise asymptomatic individual with, for example, a scan-discovered cavernous angioma, requires expert advice and involves a number of variables. If the anomaly has not bled, surgical advice might be to operate only if the headaches become worse or if positive or negative neurologic signs develop, such as seizures or paresis, respectively [8]. An AVM may or may not mandate a more aggressive approach than a cavernous angioma . Much of this decision is made between the consultant specialist (in such cases, a neurosurgeon) and the patient. The primary care clinicians and/or physiatrists should remain involved, however, because they usually get to know the particular patients better, and can help the specialist and patient to communicate, adding their own perspective, as appropriate.

Sometimes, the motivation for further workup is to protect the clinician’s legal coverage, such as when clinician and patient expectations do not coincide. It is important to be honest about this consideration, but separating social from purely medical decision making may prove difficult. We suggest integrating these points into an informed-consent approach so that the decision to undergo certain tests is a mutual agreement between patient and clinician with careful and detailed documentation. Where appropriate, evaluation by a psychologist specializing in chronic pain can prove extremely helpful in trying to decide between options.

Many patients, when diagnosed with headaches caused by an inoperable brain tumor, complications from acquired immunodeficiency syndrome (AIDS), or metastatic cancer, feel as concerned for those whom they will leave behind as they do for themselves. Spouses and partners diagnosed with terminal diseases may feel that they are abandoning their loved ones in the same way that parents’ worry about not being around for their children when they will be needed most. The person not diagnosed with the problem may feel guilty about his or her relative health. These fears often go unexpressed and may need to be addressed by caregivers to help patients, friends and family come to terms with these issues. Discussion, clarification, and resolution might also facilitate financial planning, a common concern of dying patients [9]. Some patients fear severe pain worse than death. Anxiety, depression, and suicidal ideation are common when pain is poorly controlled [10]. However, impending death in those diagnosed with a terminal disease may evoke angst and fear for reasons other than pain and may bring to consciousness spiritual concerns (not necessarily religious) that can interfere with mood and sleep. These feelings and concerns may indirectly worsen pain and suffering and increase the severity and frequency of coincident primary headaches.

In general, a head scan is obtained to rule in or out organic pathology that might account for the headache disorder. There are times when MRI, magnetic resonance angiography (MRA), magnetic resonance venography (MRV), or CT scanning is clearly the method of choice. As an introduction to the topic, some general principles might prove useful. They are discussed in greater detail elsewhere [11, 12].

CT scanning is less expensive than MRI, is usually more readily available, and takes about one-third of the time to perform. This is useful in trauma patients and in those who are delirious or have a hard time lying still. It may also be the imaging modality of choice (with thin cuts and bone windows) when calvarial tumors or skull-base pathology is suspected [13]. Except for the superiority of CT in early imaging of hemorrhage, and the clarity with which it reveals bone fractures (MRI being better for bone marrow changes), the resolution of MRI is much higher overall. MRI is not affected by the bone-reflection X-ray artifact that interferes with CT resolution at the bone–soft tissue interface; therefore, it visualizes the brain-stem and posterior fossa much better. Furthermore, MRI reveals subdural hematomas better than CT when blood is in the isodense phase with bone. MRI shows meningeal inflammation well, whereas CT scanning does not. This distinction is helpful when the patient refuses lumbar puncture, putting the patient at risk for a herniation through the foramen magnum in the skull base. It must be noted, however, that lumbar punctures have been reported to cause meningeal enhancement. A gadolinium-enhanced MRI best precedes the lumbar puncture in these cases and also replaces the need for pre-lumbar puncture CT scanning.

If the headaches are non-progressive, and stereotypic, without any sign of raised intracranial pressure or progressive neurologic dysfunction, and the neurologic examination is normal, then imaging is likely to be normal and probably not indicated. A “non-focal” neurologic examination , however, cannot rule out a midline lesion for which there may be no lateralizing signs. Examples include medulloblastomas, cerebellar astrocytomas, craniopharyngiomas, ependymomas, and tumors or cysts of the pineal region.

MRI can visualize AVMs, internal carotid dissection, sinusitis (without having to order special CT views through the sinuses), venous sinus thrombosis, and some aneurysms (although MRA or MRV may do so even better). MRA to a large degree obviates the need for more invasive cerebral angiography and can also be used to investigate the neck vessels in cases where stroke or transient ischemic attacks are of concern. MRA reliably detects aneurysms 5 mm in size or larger. It may even resolve them to 3 mm, but not as reliably as angiography. MRV is particularly useful for ruling in or out thrombosis of the venous sinuses. This is important in the differential workup of benign intracranial hypertension, especially in a potentially hypercoagulable patient who has cancer or is pregnant.

MRI is particularly useful for localizing obstruction of CSF pathways and for evaluating Arnold-Chiari malformations and lesions of the skull base. It also depicts white matter lesions associated with multiple sclerosis and small vessel disease, the differentiation between which may depend on age, presence or absence of small vessel disease, and the experience of the radiologist and clinician. Adding iodine contrast to CT, or gadolinium to MRI, markedly enhances the sensitivity of these scans to a variety of lesions. There is no cross-allergenicity between these agents, and gadolinium is safer than CT contrast in patients with compromised kidney function.

Pregnancy , the presence of metallic implants, and the types of lesions under investigation (and their expected location) affects the choice of scanning. It ·is probably best to clarify for the radiologist what is in need of being ruled in or out so that the most appropriate technology can be employed (the radiologist usually being the most up to date on evolving technology).

Situations that raise concerns about organic pathology of headache:

Reasons to obtain head scans in children with headache [14]:

When the headache history is less than 6 months or the child is under the age of 7 years, imaging should be done routinely.

The percentage of tumors that cause headache is now estimated to be lower than previously thought. This is because brain imaging for various complaints has become more common, and so-called silent tumors are increasingly being found. But, it is important to be aware that tumor-related symptoms and signs are not limited to headache and seizures. For example, sudden loss of, consciousness associated with positional changes, stroke, drop attacks, early morning nausea and vomiting with intense headache, or headache exacerbation with the Valsalva maneuver (abdominal straining) may be caused by a third ventricle colloid cyst or a pedunculated tumor blocking CSF flow. This should be evident on imaging studies. Personality changes may also be the first sign of a metastatic or primary tumor.

Headache is overestimated as a symptom of brain tumors, the location and type of which do not correlate well with location or type of tumor. Headache is a common symptom of tumors, but tumors are a rare cause of headache. Less than 1% of patients presenting to headache clinics have a brain tumor [15]. The incidence is approximately ten times less when only chronic headache is concerned. This is true at least in those who undergo head imaging despite a normal neurologic examination [16]. However, as previously noted, concern over the potential presence of a tumor may be the patient’s primary motivation for clinical evaluation. Headache as the lone symptom of a brain tumor occurs about 8% of the time [17]. The overall percentage of patients with a tumor-caused headache in neurosurgery clinics may be higher because of referral bias (the mass having often been diagnosed previously and elsewhere by brain scan).

What factors are predictive of pain? The site and rate of tumor growth may be more predictive of pain than size alone. Infratentorial and posterior fossa tumors tend to present as headache more often than supratentoria tumors [16], especially if CSF obstruction is involved. Sixty percent of childhood brain tumors are infratentorial as compared with 5–20% of adult masses. This may explain why children are more likely to present with tumor-associated headaches than are adults. Migraine with aura, even when successfully controlled with anti-migraine medication, may be tumorous in origin [18]. Furthermore, slow-growing tumors are far more likely to present with seizures than with headache [19, 20]. According to autopsy studies of patients with cancer metastases or primary intracranial tumors, leptomeningeal involvement occurs only 1–8% of the time, yet 33–76% of such patients experience headache. Other regions within the brain or cranium that are less sensitive to pain are unlikely to result in pain, unless expansion raises intracranial pressure or causes a midline shift.

Whether a tumor is primary or secondary may affect the clinical presentation. Although some literature suggests that metastatic tumors are more likely to cause headaches than are primary ones [21], other studies have found the incidence to be roughly equal [22]. Multiple sites simply make the pain less localized. Thirty percent of the time, metastatic brain tumors are the first sign of cancer anywhere in the body, but only about half of them are traced back to the primary site before death (most commonly the lung) [22]. In general, breast, lung, and melanoma cancers are most likely to invade the brain, whereas prostate cancer metastasizes to the skull, pelvis, and vertebrae.

Tumors compressing brain tissue from outside tend to induce seizures and neurologic deficits before they cause headache. Whatever the tissue type of origin, when headache occurs, metastatic or primary tumors present as tension-type headache far more often than migraine (77% vs. 9%; with 14% mixed) [23]. Migraine-like symptoms occur quite often as the result of intraventricular tumors [24]. Anti-migraine medications may occasionally alleviate the pain. This is a reason to not rely too much on the description of headache alone or response to medication as a means of reassuring patients that they do not need an MRI scan to rule out a structural cause of pain. Although most tumor-caused headaches are bilateral, there is some correlation between the most painful side and the site of the tumor. However, in cases involving considerable swelling and mass effect, false localizing signs and symptoms are common. Interestingly, raised intracranial pressure, long assumed to be a pain generator, does not seem to be the cause of pain per se [25]. It is more likely caused by displacement of, or traction on, pain sensitive structures within the cranium [2].

Most of the following medications are prescribed in the usual adult and pediatric doses. Aspirin and simple analgesics may help with mild pain. Aspirin , as well as most non-steroidal anti-inflammatory drugs (NSAIDs) should be halted if surgical intervention is likely. Aspirin irreversibly inhibits platelet function and increases the likelihood of bleeding. NSAIDs reversibly inhibit platelet function, but can still increase bleeding time. Using non-platelet affecting modified aspirin analgesics , such as trilisate or salsalate, or the new cyclooxygenase-2 (COX-2) selective NSAIDs, seems a reasonable alternative, but there are few data available to support this approach.