Two success stories in the management of Guillain–Barré syndrome illustrate the challenges of intensive care unit care in Malawi

Introduction

A disproportionate amount of the global burden of critical illness occurs in low- and middle-income countries (LMICs), where mortality attributable to these conditions remains higher than in high-income countries.^1–6 Regardless of the ability to utilise advanced therapy available in some high-resource settings, early recognition and supportive care are essential to the treatment of all critical illness. We report the intensive care unit (ICU) management of two patients with Guillain–Barré syndrome (GBS) at a referral hospital in Malawi and discuss the case-based opportunities for improving ICU medicine in low-resource contexts.

Malawi, in southern Africa, has a population of 18 million, a life expectancy of 63.8 years and a Human Development Index rank of 172/187. ⁷ It is the fourth poorest country in sub-Saharan Africa. ⁸ The prevalence of human immunodeficiency virus (HIV) in Malawi is approximately 9%, ⁹ and among adult hospital in-patients it is approximately 40%. ¹⁰ The referral hospital in the cases described is a 1000-bed public hospital in the central region of Malawi with a catchment of five million. This hospital has general wards for adult and paediatric patients, and four separate high-dependency units (HDU) for paediatric, adult, surgical and obstetric cases. It has a five-bed ICU, which serves the entire hospital. This ICU offers a 1:1 nurse-to-patient ratio, continuous non-invasive vital sign monitoring, mechanical ventilation and intravenous medication infusions. ICU care is directed by clinical officers. No physicians work full-time in the unit and there are no physician neurologists based in Malawi. Written consent was obtained from both patients for this report.

Case presentation #1

A 36-year-old man with no past medical or surgical history presented to a referral hospital in Malawi with a two-day history of progressive difficulty swallowing and breathing, and tingling in both legs (Figure 1). He reported one day of non-bloody diarrhoea approximately 10 days prior, but otherwise denied constitutional symptoms. He used no medications and reported no allergies. There was no precipitating trauma nor other trigger. He reported a negative HIV test one year prior, but did not discuss potential risk factors in the interval period. His family history was negative for any neurological or auto-immune conditions. He denied recent environmental exposures. Examination at presentation was significant for crackles over the left lower lung field, oesophageal thrush and a non-focal neurologic exam. HIV testing was positive, and CD4-testing demonstrated an absolute CD4 count of 231. Viral load testing was not performed. The patient was counseled about his new HIV diagnosis and admitted to the general ward for observation. On the ward, he developed rapidly progressive, ascending flaccid paralysis over the next two days, requiring mechanical ventilation and ICU admission on the second hospital day. Blood count at ICU admission was significant for thrombocytopenia (platelet count = 30.10⁹ per litre). Computed tomography (CT) imaging of the brain and spine revealed no lesions. Cerebrospinal fluid (CSF) testing showed no white blood cells (WBCs) or red blood cells. Further imaging and CSF studies were unavailable. By the third day, he had developed complete quadriplegia with areflexia. He was diagnosed with GBS with level 2 certainty based on the Brighton Criteria ¹¹ (Table 1). OPEN IN VIEWER

OPEN IN VIEWER

Table 1.

Brighton diagnostic criteria for Guillan–Barré syndrome. ²¹

	Level of diagnostic certainty			Case patients
Diagnostic criteria	1	2	3	Case 1	Case 2
Bilateral and flaccid weakness of limbs	+	+	+	+	+
Decreased or absent deep tendon reflexes in weak limbs	+	+	+	+	+
Monophasic course and time between onset and nadir of weakness 12 h to 28 days	+	+	+	+	+
Absence of identified alternative diagnosis	+	+	+	+	+
CSF total WBCs <50/mL (with or without CSF protein elevation above laboratory normal value)	+	± a	–	+	Unavailable
Cytoalbuminologic dissociation	+	± a	–	Unavailable	Unavailable

^aIf CSF is not collected or results are unavailable, nerve electrophysiology must be consistent with Guillain–Barré syndrome.

CSF: cerebrospinal fluid; WBC: white blood cells; mL: millilitre

The ICU treatment course included intravenous corticosteroids for 21 days. He received five days of highly active antiretroviral therapy, 14 days of prophylactic broad-spectrum antibiotics (ceftriaxone 2–4 g daily) and 12 days of antifungal treatment (fluconazole, clotrimazole). Enoxaparin for deep venous thromboembolism prophylaxis was administered for 9 days, and omeprazole for stress ulcer prophylaxis for 15 days. He remained on mechanical ventilation and underwent tracheostomy on the tenth hospital day. While in ICU, he received physiotherapy on a daily basis.

Neurological signs gradually receded in a cephalo-caudal direction, and mechanical ventilation was weaned off by day 40. He was discharged from the ICU to the general ward on day 45. Through the remainder of his hospital course, he received physiotherapy daily, and regained his ability to walk 74 days after hospital admission. After discharge, he attended weekly outpatient physiotherapy. By day 87, he had made a full neurologic recovery, reported no residual pain or neurological symptoms, and had returned to work.

Case presentation #2

A 16-month-old girl with a normal birth history, who had had isolated febrile convulsions at four months, but no surgical history, was brought to a community clinic by her mother for fever, cough and non-bloody diarrhoea (Figure 2). She had no known allergies and used no medications. Her family history was negative for neurologic or auto-immune diseases. She lived with her family in a rural village in a house with a dirt floor; her family used water from a bore hole. She had received all but one round of immunisations at the local clinic. She was prescribed a course of amoxicillin and oral rehydration salts with complete resolution of her symptoms. One week later at home, she developed a rapidly progressive ascending flaccid paralysis. She was again seen at the clinic where she was referred to the District Hospital. OPEN IN VIEWER

Upon admission, she was apyrexial (36.9℃), tachycardic (163 bpm), tachypnoeic (rate 46/min), with an oxygen saturation of 97% on room air. Examination demonstrated bilateral expiratory wheezes and weakness of all extremities. She was observed for three days, but her limb weakness worsened and progressed to respiratory distress. She was transferred to our HDU at the central hospital for supplemental oxygen. By the sixth hospital day, there was no improvement and she was transferred to the ICU for mechanical ventilation. By then, she had complete quadriplegia but normal sensation to painful stimuli. Our CT scanner was not functioning at this time, and laboratory limitations precluded CSF analysis. She tested negative for HIV and a stool sample was negative for polio. GBS was diagnosed with level 3 Brighton certainty (Table 1).

The ICU treatment included high-dose corticosteroids for 10 days. She received multiple empiric courses of ceftriaxone for presumed pneumonia. Her course was notable for multiple inadvertent extubations and aspirations of oral feeds prior to tracheostomy placement on hospital day 32. Throughout the ICU course, she received daily physiotherapy. Her signs receded gradually in a cephalo-caudal direction, and she was removed from the ventilator on hospital day 43 but remained in ICU for observation. On hospital day 56, while still in ICU, a loss of 2.5 kg in weight (from an admission weight of 9.5 kg) was recorded, stimulating continuous nasogastric feeds and ready-to-use therapeutic food (RUTF oral supplementation. She was discharged from ICU to the HDU on hospital day 75. She continued to receive physiotherapy at least three days per week in the HDU. She regained all weight lost as well as the ability to sit independently by hospital day 80, and the ability to stand on hospital day 89. She was discharged when able to walk and attended for physiotherapy from home for several months.

Discussion

These cases demonstrate not only key successes in the treatment of acute neurological critical illness in an extremely low-resource environment, but also the challenges of ICU care in this setting. Integral to these patients’ successful outcomes was early recognition and escalation of care, access to mechanical ventilation, and regular physiotherapy. Ongoing challenges in LMIC critical care medicine are illustrated by the inability to confirm a GBS diagnosis by global standards, the application of steroids for GBS and the not uncommon problem of malnutrition, acute or long-standing.

Early recognition and escalation of care is a challenge in Malawi at both community and hospital level. Most of Malawi is rural; travel to healthcare facilities often takes several days, which is costly and detracts from community duties. A 2019 study in Malawi found that 28% of children and 34% of adults presented to care >2 days after orthopedic injury. Excessive travel time contributed to the risk for delay. ¹² Other factors cited are the use of traditional healers, waiting for special prayers to prove effective before seeking care and a general perception of illness self-healing, the use of unprescribed medicine at home, and unfamiliarity with medical consultation at public institutions. ¹³ While the first case was an adult living in the city who presented directly to the referral hospital, some of these factors played a role in the case of the second patient.

Even after hospital admission, delays in the recognition of critical illness persist. Hospitals in Malawi are understaffed and clinical signs may be unnoticed or underreported. In neither of the cases reported was a full neurologic assessment of mental status, cranial nerve function, sensation, strength and reflexes performed during the hospital course. This reflects the absence of neurologic expertise and training on site. A more precise examination conducted longitudinally may have provided better insight into the disease course, particularly in the absence of more advanced diagnostic modalities. Nevertheless, respiratory failure was noted quickly in both patients because this is the trigger for ICU admission at our site. This may be controversial, but is driven by the scarcity of ventilators. Both patients were admitted to ICU immediately after developing respiratory failure. They were fortunate as our 1000-bed reference hospital has only five ICU beds available and so critically ill patients are often denied ICU admission on this account.

Access to mechanical ventilation in the ICU is but the first hurdle for critically ill patients in Malawi. In neither case was the patients’ ventilator treatment without incident (viz. unintentional extubation, ventilator-associated pneumonia, aspiration). This illustrates that ventilator management requires not only functioning ventilators themselves, but also needs protocols for ventilator settings and weaning, protocols for oral care, respiratory physiotherapists, and humidifiers. Patients who receive mechanical ventilation without these additional resources are at increased risk for complications. ¹⁴ Targeted investments should include low-cost interventions such as the implementation of simple protocols (e.g. no oral feeds for patients on mechanical ventilation) to more expensive investments such as nebulised medications. Strategic protocols for patients with respiratory failure who do not have access to mechanical ventilation may also be considered. Manual bag-mask ventilation by staff or family has been used successfully and, though demanding, is one potential strategy. ¹⁵

A simple, yet notable, part of the ICU course for both cases was the inclusion of daily physiotherapy, which was provided to both patients by a dedicated team, as well as nurses and family members. Physiotherapy has been noted for its clear benefits in ICU patients, particularly those with respiratory and neurological illness: studies have demonstrated earlier ambulation after mechanical ventilation, ¹⁶ shorter ICU and hospital stays ¹⁶ and lower one-year mortality after discharge.^17–19 Physiotherapy has great potential to improve ICU care in low-resource settings because it requires no advanced equipment and can be provided by family members. ²⁰

We cannot ignore obstacles to quality ICU care. For both patients a diagnosis of GBS was reached with lower levels of Brighton diagnostic certainty than are standard in more resourced-settings. GBS is an important public health issue in sub-Saharan Africa because of the high local prevalence of triggering tropical infections as well as HIV. ²¹ Given the substantial investment of hospital resources over nearly three months for both of our described cases and the likelihood of future admissions for GBS, clinicians in Malawi and its regional neighbours may appreciate simpler but nonetheless rigorous diagnostic criteria to make a diagnosis. This is especially important if they do not rely on invasive measurements. This may be a topic of future research.

Both our patients also received corticosteroid treatment, despite moderate quality evidence that they neither hasten recovery nor affect long-term outcome, and may according to low-quality evidence even delay recovery.^22,23 Known side-effects are insulin-requiring diabetes and hypertension. ²² Intravenous immunoglobulin and plasma exchange^23–25 are not available in Malawi, and moreover, the latter requires central catheter access (with all its attendant risks). Investments in these treatments may not be cost effective in our low-income setting, but acknowledgement of the detrimental effects of corticosteroids for GBS may be a simple first step.

Neurology training for ICU and other hospital providers in low-resource settings may improve both the ability to diagnose acute neurologic critical illness and the ability to treat it. This investment need not be a full postgraduate training in neurology; a focused training for clinical officers in the neurological conditions most frequently encountered (e.g. GBS, traumatic brain injuries, cerebral malaria) in Malawi would be most appropriate. For example, teaching ICU providers about the Brighton diagnostic criteria may increase their use and application, and so ultimately improve the clinical recognition of GBS. The next step may be the creation of treatment protocols for common neurologic illnesses, so as to avoid those treatments with undue risk (e.g. steroids). Training programmes will have the most impact if they are focused on non-physician staff members who conduct the majority of clinical care in low-resource settings.

Another obstacle was the failure to prevent acute malnutrition, a major comorbidity of prolonged ICU admission. Nutrition training in LMIC ICUs is needed: there was confusion as to whether staff members or the patient’s mother should feed the child; no clear protocol exists. There was also confusion about how to feed a patient with an endotracheal tube or a tracheostomy, which led to several gross episodes of aspiration. This was resolved after consulting the hospital dietician service, but for hospitals without dedicated dieticians, this represents a low-cost opportunity for training and improvement.

Although outside the focus of this report on acute neurologic illness, one final issue in low-resource ICU care deserves mention. The use of broad-spectrum empiric antibiotics without taking laboratory cultures is generally justified because laboratory services are unavailable or patients cannot pay for them. This of course may contribute to antimicrobial resistance.^26,27 In the cases reported, autonomic instability attributable to neurologic conditions may have been mistaken for clinical signs of infection.

A limitation to be noted in these case reports is that complete medication administration schedules were unavailable. It is nonetheless evident that successful management of neurologic critical illness is possible with low-cost interventions readily available in LMIC environments. These two cases demonstrate how early recognition, access to mechanical ventilation, and good supportive care can be life-saving. Improvements in the quality of LMIC critical care can be made by simple investments in nutrition, physiotherapy, or mechanical ventilation protocols, and by investment in appropriate neurology training as well as local research to validate feasible GBS diagnostic criteria.