The Doctor’s Advocate | Third Quarter 2012
A Diagnostic Blind Spot
And so these men of medicine consulted off and on, each gave his own opinion in both progress notes and song, though each of them was partly right, together they were wrong.
—Director’s version of John Godfrey Saxe’s 1873 poem “The Blind Men and the Elephant”
A 34-year-old morbidly obese female with a history of hypertension, type 2 diabetes, and depression underwent gastric bypass surgery on September 21. She weighed 402 pounds and was 5 feet 6 inches tall. Surgery was uneventful, and she was discharged on September 23.
Two days later, she presented to the ER with epigastric pain; she was placed on a clear liquid diet, prescribed Reglan, and sent home. She was readmitted on October 22 for vomiting, abdominal pain, and dehydration. An upper GI series revealed gastric outlet obstruction. She underwent revision of the gastrojejunostomy and was discharged on October 29, but was readmitted three days later for vomiting with epigastric pain. She was placed on a clear liquid diet and discharged in three days.
She was readmitted on November 9 for hyperemesis and malnutrition and weighed 356 pounds. She received nutritional support with daily multivitamin and mineral supplements. A neurological examination by the surgeon was normal. On November 27, she experienced hallucinations with blurred vision, dizziness, and headaches; her blood pressure was 151/103. A psychiatrist’s impression was “acute episode secondary to a metabolic neurological event with underlying psychological issues.” The following evening, she insisted there were children under her bed. She was given Haldol and Ativan, and the episode resolved.
On December 1, neurologists were consulted. They and their nurse practitioner (NP) briefly discussed the possibility of Wernicke’s encephalopathy (WE) but concluded the patient was suffering from psychological, not neurological, issues because their neurological exam was normal and the classic signs and symptoms of WE (nystagmus, ataxia, and confusion) were absent.
The NP thought the symptoms could have a metabolic etiology and ordered an MRI and EEG plus blood folate, vitamin B12, vitamin D, and homocysteine levels. All were normal except the folate, which was indeterminate. The MRI indicated a symmetric T2 abnormality in the posterior medial thalamus and associated restricted diffusion. The report stated the findings might be due to venous ischemia, a metabolic disturbance, or encephalitis. Acute dural venous sinus thrombosis was ruled out by a normal MRV. The EEG showed frequent generalized delta slowing consistent with generalized cerebral dysfunction. Suggested etiologies included toxic and metabolic conditions. The NP concluded that the patient was suffering from an encephalopathy with low folate and ordered folic acid supplements.
On December 4, the patient became depressed and started crying. The NP recommended antidepressants with psychiatric counseling and ordered Celexa 40 mg daily. The psychiatrist opined this was an acute psychotic episode of depression, etiology undetermined. An ophthalmologist was consulted to evaluate right eye pain and noted nystagmus, possibly secondary to medications. Hyperalimentation/total parenteral nutrition (HAL/TPN) was ordered. On December 7, the psychiatrist noted the patient was intermittently confused and mildly agitated and increased maintenance Haldol. A physical therapy evaluation on December 10 documented vertical nystagmus and charted that the neurologists and nurses had been notified. In an infectious disease consultation for a urinary tract infection (UTI), nystagmus was again noted.
On December 14, the surgeon directed the nurses to wean the patient off HAL/TPN and have metabolic support services instruct her on a bariatric diet. The pharmacy’s parenteral nutrition fluid order form dated December 14 indicated that 500 mg of thiamine should be added to the base formula to treat possible thiamine deficiency. However, the surgeon canceled this order because he thought one 500 mg dose of thiamine had already been given before he wrote the order to discontinue the HAL/TPN. On December 15, the physical therapist again noted nystagmus and that the patient was lethargic and unable to stand.
On December 16, the patient was found sleeping on the floor and did not know how she got there; the psychiatrist decreased her Ativan. On December 18, the HAL/TPN was discontinued, and the internist who managed her hypertension documented horizontal nystagmus. A neurological examination by the neurologists showed bilateral lower and upper extremity strength, no focal deficits, and symmetrical deep tendon reflexes.
On December 23, the patient was lethargic, and the physical therapist documented nystagmus and ataxia. On December 26, she developed fever due to her UTI, and the neurologists’ impression was encephalopathy due to an infectious process, depression, and psychosis. They did not feel further neurological workup was necessary because all tests were normal. On January 18, the patient again exhibited confusion and lethargy; the neurologists found no neurological deficits and concluded that her symptoms were secondary to a metabolic or infectious process.
In late January, the nutritionist charted that the patient was not meeting her nutritional needs, and she underwent surgical placement of a feeding jejunostomy. She was then transferred to a long-term care facility where the feeding tube ruptured. She was taken to the university hospital where doctors diagnosed WE and treated her with thiamine. Shortly thereafter, she began to recognize people and was able to walk, brush her teeth, and dress and feed herself.
All bariatric surgery, neurology, and internal medicine experts opined that failing to diagnose WE in a post–gastric bypass patient, with hyperemesis, who displayed the classic signs of confusion, nystagmus, and ataxia with an MRI abnormality in the medial thalamus was below the standard of care. They were critical of the medical team for failing to seriously consider thiamine deficiency, because it is a well-recognized complication of gastric bypass surgery, and whenever a clinician suspects a metabolic disorder in this setting, thiamine should be given.
Early in her hospitalization, the patient exhibited psychological issues with signs of depression, which may have prevented her physicians from recognizing that the onset of new symptoms was a psychosis related to WE. However, once the MRI report was received and dural venous thrombosis ruled out, thiamine should have been administered. The patient, who failed to improve despite being in the hospital for a long time, should have been seen more often by all specialists. The neurologists saw the patient only four times from December 1 through mid-February, and while their neurological examinations were “normal,” they failed to note the nystagmus and ataxia, which were well documented in the chart.
There was failure to establish a differential diagnosis. Although the neurologists considered metabolic issues and prescribed folate, they did not prescribe thiamine. The neurologists, bariatric surgeon, radiologist, psychiatrist, and ophthalmologist all had pieces of information that should have led to the correct diagnosis, but they communicated poorly with one another, and no one took the responsibility to put everything together. All caregivers were criticized for not being more aware that the patient was not meeting her nutritional needs. Without expert support, the claim was settled.
Wernicke’s encephalopathy is a neurologic complication of thiamine (vitamin B1
) deficiency. It is characterized by the classic triad of mental confusion, oculomotor dysfunction, and gait ataxia. Punctate hemorrhages around the third and fourth ventricles are characteristic. While most often associated with chronic alcoholism, in one autopsy series, non–alcohol abusers accounted for 23 percent of cases. Other conditions associated with WE include anorexia nervosa, prolonged intravenous feeding without proper supplementation, prolonged fasting or starvation, dialysis, and gastrointestinal surgery (especially bariatric surgery).
The classic triad is recognized in only one-third of patients. In one study, mental status abnormalities were present in 82 percent, ataxia in 23 percent, oculomotor abnormalities in 29 percent, and polyneuropathy in 11 percent. The encephalopathy is characterized by confusion, disorientation, indifference, and inattentiveness. The oculomotor dysfunction includes nystagmus and lateral rectus palsy. Nystagmus is the most common finding and is typically evoked by horizontal gaze; vertical nystagmus can also occur. Ataxia primarily involves stance with a wide-based gait and slow, short-spaced steps.
WE is primarily a clinical diagnosis. Laboratory and neuroimaging studies are helpful, but the biggest barrier to diagnosis is a low index of suspicion. While there are no diagnostic laboratory tests, measurement of serum thiamine or thiamine pyrophosphate may be useful. However, results from these tests are not necessary for patient management, and whenever WE is considered in the differential diagnosis, immediate thiamine replacement is indicated. Only half of patients demonstrate EEG abnormalities, usually diffuse slow wave activity. CT may show symmetric, low-density abnormalities in the diencephalon, midbrain, and periventricular regions. However, a normal CT scan does not rule out the diagnosis. MRI is more sensitive than CT, and typical findings include areas of increased T2 and decreased T1 signal surrounding the aqueduct and third ventricle and within the medial thalamus and mammillary bodies.
Treatment with intravenous thiamine is safe, inexpensive, and effective. Diagnostic testing should not delay treatment, which should immediately follow consideration of the diagnosis. A recommended regimen is 500 mg of thiamine intravenously, infused over 30 minutes, three times daily for two consecutive days and 500 mg intravenously or intramuscularly once daily for an additional five days. Oral administration of thiamine is an unreliable initial treatment.
WE is a well-known complication of bariatric surgery and typically occurs four to 12 weeks following surgery; female patients with vomiting appear to be at higher risk. Since WE may be precipitated by glucose loading in patients with unsuspected thiamine deficiency, thiamine should be administered prior to or along with glucose infusion in patients who are at risk for thiamine deficiency.