Weakness, Abdominal and Limb Pain in a 17-Year-Old With Type 1 Diabetes

Abstract

A previously healthy 17-year-old male presented to the emergency department (ER) with a month-long history of widespread abdominal pain, non-bilious vomiting and early satiety, notably 3 months after a new diagnosis of type 1 diabetes mellitus (T1DM). He was afebrile with no history of bloody stools, and a negative review of constitutional, genitourinary, cardiovascular, respiratory and neurological systems. A recent CT of his abdomen showed no evidence of mechanical obstruction. This had been the patient's fourth ER visit for abdominal pain in the last 2 months. Since his T1DM diagnosis, he had been adherent to his insulin regimen, and his most recent glycosylated haemoglobin (HbA1c) was 8.5%, a reduction from 13.6% at diagnosis. Upon admission, his point-of-care glucose was elevated, however, blood gas analysis was within normal values, and there was no biochemical evidence to support a diagnosis of diabetic ketoacidosis. Whilst in the hospital, gastroenterology was consulted, and a scintigraphic gastric emptying test was performed, confirming delayed gastric emptying. However, despite treatment with domperidone, his abdominal pain remained refractory to best medical management, and he developed several new symptoms. He began to complain of a symmetrical burning pain, followed by paresthesia, in his ankles, gradually ascending to his mid-calves, with associated objective weakness in his feet, ankles and knees. He was found to have several autonomic symptoms, including systolic hypertension ranging from 140–150/90–110 mmHg, with a tachycardia of 100–120 bpm. His neurologic review of symptoms was negative for focal neurological signs and symptoms, headaches, vision loss, cranial neuropathies and bowel or bladder dysfunction.

On exam, the patient had normal function of his cranial nerves, diminished power of 4/5 along his feet, ankles and knees bilaterally, +1 reflexes at the knees and absent reflexes at the ankles. In the lower extremities, there was a diminished pinprick, temperature, pain and vibration sensation below the knees with significant allodynia. His neurologic exam in his upper extremities was normal. He had normal pulmonary effort, a normal cardiovascular exam, a normal abdominal exam and no lymphadenopathy.

There was no family history of autoimmune diseases or neurological disorders.

The patient's blood work was within normal limits, with normal haemoglobin concentration (160 g/L), white count (7.0 × 10E9/L), neutrophils (4.7 × 10E9/L) and pH (7.38). Biochemical markers were normal, including creatinine (75 μmol/L), urea (4.4 mmol/L), and thyroid hormone (3.69 mIU/L), with negative urine ketones. Further investigations, including vitamin B12 (456 pmol/L) and vitamin E (alpha tocopherol 11.5 μmol/L, gamma tocopherol 2 μmol/L) were within normal ranges. A lumbar puncture showed elevated protein (0.76), elevated albumin (0.511), with a normal nucleated cell count (1.0) and negative CSF inflammatory markers in keeping with albuminocytologic dissociation. An MRI of the brain and spine was negative for nerve root enhancement or any other abnormalities.

Given the elevated CSF protein, the patient was treated with IVIG for suspected acute demyelinating inflammatory polyradiculopathy (ADIP), also known as Guillain-Barre syndrome. However, the patient's symptoms continued to worsen, most notably with significant burning limb pain. His hypertension was minimally responsive to hydralazine. Given an unremarkable urinalysis, urine catecholamines, and renal ultrasound, his elevated blood pressure was suspected to be secondary to pain. The pain was poorly responsive to first-line neuropathic agents, including gabapentin and nortriptyline, and showed no significant improvement with opioid therapy, including morphine. This clinical suspicion was confirmed as his blood pressure later stabilised to 130/70s with improved pain control.

The patient required transfer to a tertiary care centre for further investigations and management. A nerve conduction study revealed a sensorimotor axonal length-dependent polyneuropathy, which combined with the clinical picture and lack of response to typical ADIP treatment, was suggestive of Treatment-induced neuropathy of diabetes (TIND).

TIND, previously known as insulin neuritis, is primarily a small fibre neuropathy induced by aggressive glucose-lowering pharmacological or lifestyle measures in patients with a history of poor glycemic control [1]. The pathophysiology is not fully understood but is thought to result from endoneural ischemia and regeneration caused by rapid glucose deprivation, resulting in apoptosis [2]. This can be particularly evident in patients with a prolonged history of hyperglycemia, in whom a rapid transition to a euglycemic state may impose biochemical changes and metabolic stress, resulting in neuronal injury.

The clinical manifestations of TIND include acute-onset neuropathic pain affecting small fibre sensory modalities [1, 2]. Symptoms are commonly accompanied by allodynia, hyperalgesia and autonomic symptoms such as orthostatic hypo- or hypertension and gastrointestinal disturbances such as nausea and vomiting [2]. Furthermore symptoms of TIND, as opposed to generalised diabetic neuropathy, also differ in that they are reversible for some patients [3].

The severity of the neuropathy has been reported to correlate with the magnitude of the percent change in the HbA1c [2]. Current observational data reports neuropathy can be triggered by a greater than 2% reduction in HbA1c over 3 months [1]. The diagnosis of TIND is predominantly clinical, relying on symptomatology, a drop in HbA1c greater than 2% in 3 months or longer, and a history of onset following the HbA1c reduction, with supporting investigations such as electroneuromyography or skin biopsy [1].

Those at risk are patients with poorly controlled diabetes, more commonly in Type 1 diabetics, with a high initial A1c treated with aggressive glucose-lowering strategies [1]. Notably, an absolute risk of TIND has been reported at approximately 20% with a reduction in HbA1c exceeding 2% and greater than 80% when the decrease surpasses 4% (4). Studies have also indicated that females and those with eating disorders were an at-risk subgroup for developing TIND [1]. Although regarded as a rare consequence of rapid glycaemic correction, a retrospective study identified the prevalence of TIND in up to 11% of patients evaluated for diabetic neuropathy. Whilst paediatric data is limited, cases have been documented, suggesting that TIND can also occur in children (3). TIND has not been studied extensively, and therefore, a standardised treatment has not been established [1]. Currently, management includes a multidisciplinary team approach to provide adequate glycaemic control and supportive treatment. In the short term, painful neuropathy has been shown to improve by aiming for higher glycaemic targets. However, gradual and long-term optimisation of blood sugars should be prioritised, as near-complete functional recovery was more likely in patients with stable glucose levels compared to those with unstable glycaemic control [2]. Referral to a neurologist should be included to ensure appropriate neurological workup and management of the painful neuropathy [1]. Along with the augmentation of glucose control, neuropathic agents can be used to alleviate pain, with recommendations against using opioids as they have limited benefit for neuropathic pain and can worsen gastrointestinal symptoms. In patients with persistent GI symptoms, a dietician should be involved in dietary modifications [1]. Furthermore, a gastroenterology referral should also be obtained to rule out autonomic dysfunction of the gastrointestinal tract, using gastric emptying scintigraphy, as it is the gold standard for the diagnosis of gastroparesis [1]. Finally, follow-up is required for preventative screening of further autonomic neuropathy, including an annual retinopathy exam and an annual urine ACR to rule out nephropathy [1, 2].

The authors declare that this case report did not require an ethics committee or institutional review board. All procedures involving human participants were conducted in accordance with the ethical standards of Queen's University, Ontario, Canada, and the Declaration of Helsinki (1964), as revised in 2013.

Written consent was obtained from the patient and his family for publication of this case.

The authors declare no conflicts of interest.