The practical nature of neuropsychiatric practice usually requires the clinician to offer a diagnosis found in the Diagnostic and Statistical Manual of Mental Disorders.¹ Given the realistic limitations of neuroscience as it applies to the neurological basis of psychiatric illness, the clinician is often challenged to provide an accurate diagnosis based on DSM-IV-TR terminology. This article explores the issues encountered by neuropsychiatrists regarding proper diagnosis based on the criteria established in the DSM-IV-TR. It is our hope that these issues will be addressed in the new DSM-5.

The DSM is a useful reference that relies on empirical foundations to identify discrete psychiatric disorders. It is designed to be a helpful guide to clinicians, and it provides sets of clinical criteria and statements of the constructs embodied in the clinical criteria. Each section in the DSM-IV-TR includes diagnostic features; associated features and disorders; specific cultural, age, and gender features; prevalence, course, familial pattern, differential diagnosis, recording procedures, and diagnostic criteria.

The DSM-IV-TR section on Disorders of Infancy, Childhood, or Adolescence defines several conditions that could be confidently identified as neuropsychiatric. These include Pervasive Developmental Disorders (299.X), Tic Disorders (307.2X), Mental Retardation (317, 318.X), Learning Disorders (315.X), Motor Skills Disorder (315.4), Communication Disorders (307.X, 315.3X), and Attention-Deficit and Disruptive Behavior Disorders (312.X, 313.81, 314.X).

From the perspective of the neuropsychiatrist, the DSM-IV-TR classification for disorders of infancy, childhood, and adolescence are generally comprehensive in scope and reflect the connection between abnormal neurological development and the presence of neuropsychiatric signs and symptoms. They present a very good example of a neuropsychiatric nosology. The text identifies useful elements of neurological and neuropsychological evaluation and identifies associated neurological signs and symptoms. This section of the DSM-IV-TR represents an ideal template for adult neuropsychiatric disorders.

Unfortunately, ideal templates for adult neuropsychiatric disorders are lacking in DSM-IV-TR. Many syndromes encountered in neuropsychiatric practice are not identified in the DSM-IV-TR. The diagnostic criteria often rely on the clinician to definitively link the psychiatric symptoms to concurrent neurologic conditions in the absence of conclusive neuroscientific proof. Furthermore, the diagnostic nomenclature is often lengthy and clumsy.

This article will review several common neuropsychiatric disorders and syndromes as seen in clinical practice and illustrate the challenges and limitations of DSM-IV-TR criteria.

I find the DSM-IV-TR section on Delirium (293.0) to be a helpful clinical and diagnostic guide. It provides useful clinical examples to assist the clinician in recognizing the varied presentations of delirium. Meagher et al.² have proposed improving the DSM-5 classification of delirium by expanding the diagnosis to include subtypes “With Hyperactive Features” (Table 1) “With Hypoactive Features” (Table 2), and “With Mixed Hyperactive/Hypoactive Features.”

Caregivers of patients with delirium are often confused by the noteworthy differences in activity levels manifested by the patient. Consequently, patients with diminished activity are frequently overlooked as having delirium. Acknowledging the existence of varied presentations of delirium would improve our ability to educate caregivers, such as nursing staff in the general hospital, and thus improve detection and treatment of delirium.

The DSM-IV-TR section on Cognitive Disorders is problematic. The “A” diagnostic criteria for Dementia of the Alzheimer Type (294.1X) represent the prototype for all specified DSM-IV-TR dementia diagnoses (Table 3). These include Vascular Dementia (290.4X), Dementia Due to Other General Medical Conditions (294.1X), and Substance-Induced Persisting Dementia (292.82); yet many well-established dementia syndromes do not present in this way, for example the Progressive Aphasias or Posterior Cortical Dementia. Consequently, the diagnostic nomenclature for DSM-IV-TR dementias departs from the typical terminology utilized by clinicians.

Table 4 illustrates DSM-IV-TR terminology for some of the more common dementias. It clearly demonstrates the limitations of the DSM-IV-TR diagnostic terminology for diagnosing dementias in clinical practice. I believe the criteria would be improved by defining the specific presentation of the various dementias as “syndromes” and providing consensus criteria for each syndrome in the DSM-5. The best current illustration for this would be the McKeith criteria³ for Dementia with Lewy Bodies (which would be reclassified as the Dementia With Lewy Body Syndrome) and the DSM-IV-TR criteria for Dementia of the Alzheimer Type (which would be reclassified as the Dementia Syndrome of the Alzheimer Type).

In this way, we can align the DSM-5 criteria with the typical nomenclature used by clinicians and also eliminate the conundrum of certainty that is required for dementia diagnoses in the DSM-IV-TR because, by definition, syndromes are not diseases. For example, clinicopathologic studies of dementia have demonstrated that the correlation between bedside diagnosis and histological findings ranges widely, from 65% to 95%.⁴ Londos et al.⁵ found that two-thirds of patients carefully diagnosed with probable Alzheimer’s disease (AD) by NINCDS-ADRDA criteria⁶ also met McKeith criteria³ for Dementia with Lewy Bodies (DLB). Furthermore, the neuropathology for each group of patients was remarkably similar. Both the AD and the DLB groups had histologically-identified neuritic plaques, neurofibrillary tangles, and Lewy bodies. This problem with diagnostic specificity is thoroughly explored by Dubois et al.⁴

In reality, the diseases that cause dementia syndromes are primarily limited to some combination of amyloidoses causing neuritic plaques or vascular lesions, tauopathies causing neurofibrillary tangles, synucleinopathies causing Lewy bodies, prionopathies causing spongiform changes, or cerebrovascular disease causing strokes or hemorrhages. Each of these diseases can present with any number of dementia syndromes, depending solely on the location of the pathology, as opposed to the specific type of histopathology. Thus, each dementia syndrome can reflect different underlying diseases.

The DSM-IV-TR criteria for vascular dementia are particularly problematic because they rely strongly on clinical judgment to decide whether the patient’s neurological signs and symptoms are “etiologically related to the disturbance” and are thus causative. I have seen many cases where patients are given the diagnosis of vascular dementia simply because they have radiologic evidence of cortical, white-matter, or lacunar infarcts, regardless of the location in relation to the cognitive findings. Furthermore, many of these patients have cortical atrophy, suggesting the presence of a cortical degenerative component. The certainty of identifying a stroke on neuroimaging usually exceeds the certainty that atrophy is disease-related rather than age-related. Thus, we will tend to overdiagnose vascular dementia at the expense of what is more likely a mixed cortical degenerative disease with a vascular component. Wallin et al.⁷ have proposed a more detailed and clinically useful set of criteria for vascular dementia by defining subtypes (Table 5).

DSM-IV-TR coding procedures for neuropsychiatric syndromes in the setting of dementia are inconsistent. Vascular Dementia (290.4) is the only type of dementia that has defined subtypes to indicate the presence of significant associated symptoms, such as delirium (290.41), delusions (290.42), or depressed mood (290.43). Any associated behavioral disturbance would be specified, but not coded. On the other hand, Dementia of the Alzheimer Type (290.1X) with associated behavioral disturbances would be coded as 290.11. But there would be no coding for an associated delirium, delusions, or depressed mood. This is illustrated in Table 6, where the DSM-IV-TR diagnoses for the very common presentation of an elderly patient with dementia, delusions, depressed mood, and intermittent aggression are provided. The DSM-5 criteria would be improved by standardizing these coding procedures.

Finally, the DSM-IV-TR section on the dementias offers proposals for the office examination of various cognitive domains. Coffey et al.⁸ identified excessive variation in the actual office assessment of cognition by clinicians who describe their practice as neuropsychiatric. Furthermore, the Institute of Medicine has identified variation in clinical practice as a significant impediment to achieving higher-quality care.⁹ Ideally, we can agree on a single, valid method for office evaluation of each cognitive domain and specify these tests in the DSM-5. The diagnostic criteria for each dementia syndrome could specifically include the findings on each test. For example, diagnosing Frontotemporal Dementia Syndrome would require the clinician to administer an agreed-upon office test of executive cognition and identify a cutoff score that would serve as one means of qualifying the patient for the diagnosis. In this way, we could eliminate the unacceptable variations with interrater reliability and ensure more consistency in diagnosing these often-confusing syndromes.

Neuropsychiatrists typically treat patients whose psychiatric symptoms appear in the setting of general medical conditions, such as traumatic brain injury, autoimmune disorders, striatal disorders, and developmental disorders. The DSM-IV-TR includes diagnostic criteria for these clinical presentations, which include psychotic disorders, mood disorders, anxiety disorders, personality changes, and cognitive impairment (Table 7). In order to offer the diagnosis, the clinician must conclude that the general medical condition is actually causing the psychiatric symptom. This degree of certainty is frequently not possible, given our current understanding of neuroscience and functional neuroanatomy. Bradford-Hill proposed specific criteria to assist the clinician in identifying causation.¹⁰

Even under the best clinical circumstances, it is often not really possible to conclude that the general medical condition is causative. For example, a 50-year-old man develops a manic syndrome a few days after suffering a right frontal cortical infarct. His father and brother have been diagnosed with bipolar disorder, but the patient has never had a major mood disorder. Is the manic syndrome a sequela of right-hemisphere stroke (293.83 Mood Disorder Due to Stroke With Manic Features) or the new onset of mania in a man who is genetically predisposed and now experiencing significant stress because of his disability (296.4 Bipolar Disorder, Type I, Most Recent Episode Manic)?

In the absence of truly consistent and reproducible evidence of neuropsychiatric causation in many cases, I believe it is most reasonable at present to eliminate the need for certainty. When the psychiatric syndrome occurs in the setting of a known condition such as traumatic brain injury, stroke, or dementia, then the clinician could simply identify the DSM-IV-TR subtype (Table 7) and diagnose the patient with, for example, Right Hemisphere Stroke With Manic Features. This format would leave open the possibility of causation, but not insist on it, while acknowledging that there is a larger biologic context for the mood disorder in the patient.

Apathy is a common neuropsychiatric syndrome that is not clearly identified in the DSM-IV-TR. Indeed, an OVID database search done in late 2010 revealed over 2,000 citations for the keyword “apathy.” It is often quite disabling to the patient and frustrating to caregivers who may confuse the apathy with depression or volitional, but undesirable, behavior. Starkstein et al.¹¹ and Marin et al.¹² have proposed that apathy syndromes be formally identified in the DSM (Table 8). Apathy can occur in various contexts, such as poststroke or frontotemporal dementia. Currently, apathy in these contexts would be diagnosed as 310.1 Personality Change Due to XXX, Apathetic Type. However, developing formal consensus criteria would support reliable case description and identification, facilitate clinical communication about the syndrome, and encourage research and treatment efforts.

In addition to apathy, there are other common neuropsychiatric syndromes not clearly defined in the DSM-IV-TR. These are noted in Table 9.

The evolution of the DSM has reflected advances in the biological understanding of mental illness, and it also mirrors changes in our culture and how we perceive mental illness. The formal diagnostic criteria provided in the manual permit reliable diagnosis, which guides evidence-based treatment, epidemiologic studies, and clinical research. The DSM-IV-TR is, however, less useful to neuropsychiatrists for everyday clinical work. Most neuropsychiatric syndromes occur in the context of neurological and general medical illness, yet conclusive proof of causation is often lacking. Furthermore, the neuropsychiatric lexicon offered by the DSM-IV-TR is rather clumsy and often requires multiple diagnoses and codes (Table 6) to accurately diagnose common neuropsychiatric clinical presentations. It is my hope that these issues will be addressed in the DSM-5.