A 73-year-old Nepalese woman was admitted to the medical ward of Ruttonjee Hospital on 13 June 2020. She enjoyed good past health and worked as a street cleaner. She presented with a 3-day history of fever, headache, and nausea. On admission, she was conscious and alert, and physical examination was unremarkable. Computed tomography (CT) of the brain was likewise unremarkable. Leukocytosis (white cell count 12.73×10⁹/L) with raised neutrophils (absolute neutrophil count 7.43×10⁹/L) and raised eosinophils (1.91×10⁹/L) was noted.

Lumbar puncture yielded clear cerebrospinal fluid (CSF) with raised opening pressure of 29 cm H₂O. Cerebrospinal fluid showed leukocytosis (CSF white cell count 510/mm³, lymphocytes: 76%), low glucose (CSF: 2.1 mmol/L; serum: 5.9 mmol/L) and high protein levels (1.24 g/L) but tested negative for bacteria, virus, acid-fast bacilli smear, fungal smear, and cryptococcus antigen. Empirical treatment for tuberculous meningitis including isoniazid, rifampicin, ethambutol, pyrazinamide, and dexamethasone was commenced. She responded quickly with resolution of fever and headache.

Nonetheless her condition deteriorated 5 days later with recurrence of fever, new-onset confusion, vomiting, and urinary incontinence. Alanine aminotransferase increased to 368 IU/L. In view of her deranged liver function, tuberculosis treatment was stopped. Serial eosinophil count rose from 1.91×10⁹/L to 4.97×10⁹/L. Plain and contrast brain CT were unremarkable with normal ventricle sizes. Magnetic resonance imaging of the brain showed inflammatory changes in bilateral cerebral and cerebellar sulcal spaces and the leptomeningeal region.

Due to new-onset confusion, incontinence and fever, a second lumbar puncture was performed and revealed a high opening pressure of 21 cm H₂O, turbid CSF with a rising white cell count of 920/mm³ (lymphocytes: 60%, polymorphs: 40%), low glucose (CSF: 2.3 mmol/L; serum: 6.0 mmol/L) and high protein levels (0.71 g/L). Eosinophilic meningoencephalitis was suspected and a pathologist was consulted for eosinophil screening of her CSF. Eosinophils were found. The patient had no rash, diarrhoea, drug history or allergy that could account for eosinophilia. Stool analysis for parasites, ova and cysts was negative. Autoimmune markers including antineutrophil cytoplasmic antibody were negative. Cerebrospinal fluid did not reveal malignant cells. Cerebrospinal fluid was sent for polymerase chain reaction testing for parasites and was found to be positive for Angiostrongylus spp.

A further food history was explored. The patient had ingested raw snails from an unknown source 2 weeks prior to admission. She had a habit of consuming raw snails to boost bone health, a Nepalese custom.

She was prescribed prednisolone at a dose of 40 mg daily for 2 weeks then tapered off. Peripheral eosinophilia resolved within 5 days. She made a full neurological recovery and was discharged home.

Angiostrongylus cantonensis is a parasitic helminth. It was first described in ‘Canton’ Guangzhou, China in 1935. Rats are the primary hosts. Eggs hatch in the rats’ lungs, and first-stage larvae (L1) are passed in their faeces. Snails and slugs that feed on rat faeces act as intermediate hosts (L2). Larvae reach the third stage (L3) in the intermediate host and can infect humans when they consume slugs, snails or contaminated vegetables.1

Following human host ingestion, larvae (L3) are released in the gastrointestinal tract where they enter the bloodstream. They penetrate the liver, brain and spinal cord where they moult to mature stages (L4 and L5). Most cease to develop and die within the central nervous system. Nonetheless the presence of larvae alone attracts lymphocytes, plasma cells and eosinophils, generating an inflammatory response.

Since Angiostrongylus was first reported in 1945, over 2000 cases have been reported worldwide, mostly in Asia (Thailand: 47%; China: 27%), while four cases have been reported in Hong Kong.2 It is the most common parasitic cause of eosinophilic meningitis. Symptoms are often mild or moderate and include headache, vomiting, photophobia, and neck stiffness. Fever is uncommon. In severe cases, heavy infestation can lead to coma and death.3

Angiostrongylus meningitis infection typically gives a CSF finding of leukocytosis with eosinophilia, and raised protein level with normal to low glucose level.1 With the exception of eosinophilia in CSF, this mimics tuberculous meningitis.

The degree and evolution of eosinophilia can be variable. A study of infected travellers to Jamaica observed that although all their patients developed eosinophilia at some point, fewer than half had peripheral eosinophilia on their initial blood test. Eosinophil level did not peak until 2 weeks later. Only half of affected patients had eosinophils in their CSF at the time of their first lumbar puncture.4 In our patient, eosinophils were not found in the first lumbar puncture although the peripheral eosinophil count continued to rise and did not peak for 13 days.

Causes of eosinophils in CSF include parasites (angiostrongyliasis, gnathostomiasis, and baylisascariasis), other infectious agents (bacteria, fungal, and virus), haematological malignancy, and adverse drug reactions. Diagnosis is often difficult. Larvae occur in only a number in the CSF and are hard to identify. Serological methods such as enzyme-linked immunosorbent assay and immunoblotting are not often available and are limited by cross-reactivity with other helminths.3 Polymerase chain reaction of Angiostrongylus cantonensis DNA in the CSF is the most accurate way to confirm a diagnosis with 70% to 80% sensitivity and 99% specificity.3

Choice of anthelmintic drugs is controversial as there have been few clinical trials. They are more effective when started early, preferably within 2 weeks of infection, as they work better against younger larvae. By the time patients seek medical attention, larvae in the central nervous system have reached maturation, rendering treatment ineffective. Another concern is that worm death caused by anthelmintic drugs may trigger a severe inflammatory reaction and aggravate neurological symptoms.5

Prednisolone has been shown to be superior to placebo in reducing headache duration. Corticosteroids may reduce intracranial pressure and blunt the inflammatory reaction to dying worms.6 Lumbar punctures are also reported to be effective in reducing intracranial pressure and relieving headache.3

There was a seemingly stable period after our patient’s first lumbar puncture and the commencement of tuberculous meningitis treatment, with subsidence of fever and headache. Her transient improvement may have been due to the use of dexamethasone (as part of tuberculous meningitis treatment) and the relief of intracranial pressure by lumbar puncture.

The present case demonstrates that clinicians should be aware of the differential diagnosis of Angiostrongylus meningoencephalitis, especially in the context of eosinophilia, snail ingestion or travel to endemic areas. Timely confirmation by polymerase chain reaction testing of CSF is important. Supportive therapy with corticosteroids may provide symptom relief.

Both authors contributed to the concept of study, acquisition and analysis of data, drafting of the manuscript, and critical revision of the manuscript for important intellectual content. Both authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

Both authors have disclosed no conflicts of interest.

This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

The patient was treated in accordance with the Declaration of Helsinki and verbally agreed to publication of this anonymous case report.

1. Ansdell V, Wattanagoon Y. Angiostrongylus cantonensis in travelers: clinical manifestations, diagnosis, and treatment. Curr Opin Infect Dis 2018;31:399-408. Crossref

2. Wang QP, Lai DH, Zhu XQ, Chen XG, Lun ZR. Human angiostrongyliasis. Lancet Infect Dis 2008;8:621-30. Crossref

3. Murphy GS, Johnson S. Clinical aspects of eosinophilic meningitis and meningoencephalitis caused by Angiostrongylus cantonensis, the rat lungworm. Hawaii J Med Public Health 2013;72(6 Suppl 2):35-40.

4. Slom TJ, Cortese MM, Gerber SI, et al. An outbreak of eosinophilic meningitis caused by Angiostrongylus cantonensis in travelers returning from the Caribbean. N Engl J Med 2002;346:668-75. Crossref

5. Hidelaratchi MD, Riffsy MT, Wijesekera JC. A case of eosinophilic meningitis following monitor lizard meat consumption, exacerbated by anthelminthics. Ceylon Med J 2005;50:84-6. Crossref

6. Chotmongkol V, Sawanyawisuth K, Thavornpitak Y. Corticosteroid treatment of eosinophilic meningitis. Clin Infect Dis 2000;31:660-2. Crossref

An 8-year-old girl with severe aplastic anaemia and failed immunosuppressive therapy underwent a matched-sibling bone marrow transplantation. Neutrophils and platelets were engrafted 23 and 27 days after transplantation, respectively. Regeneration of marrow and full donor chimerism were demonstrated 30 days after transplantation. Progressive liver dysfunction and cholestasis were noted 5 weeks post-transplantation with peak alanine aminotransferase level of 878 IU/L (reference value, <35), aspartate aminotransferase level 966 IU/L (reference range, 10-40), gamma-glutamyltransferase level 2065 IU/L (reference range, 13-28), total bilirubin level 445 μmol/L (reference range, 10-24), and direct bilirubin level 430 μmol/L (reference range, 5-10). Cholesterol levels were also elevated with total cholesterol level of 23.1 mmol/L (reference value, <5.2), high-density lipoprotein-cholesterol (HDL-C) level 0.5 mmol/L (reference value, >1.6), and high triglycerides (TG) level 11.8 mmol/L (reference value, <1.7). Low-density lipoprotein cholesterol (LDL-C) level could not be calculated based on indirect quantitation with the Friedewald equation as per usual practice since TG level was >4.5 mmol/L; hence, it was measured directly and was normal at 0.2 mmol/L (reference value, <4.1).

Apolipoprotein (Apo) A1 level was low at 0.38 g/L (reference range, 1.2-2.0) and Apo B level was elevated to 1.88 g/L (reference range, 0.41-1.07). Lipoprotein electrophoresis showed chylomicron and very low–density lipoprotein bands, and an additional beta-lipoprotein band that had migrated to cathode was detected, compatible with lipoprotein-X (Lp-X) [Fig a]. There was no family history of hypercholesterolaemia and clinical examination did not reveal any xanthoma. With improvement of cholestasis, dyslipidaemia gradually resolved by 2 years post-transplant with expectant management (Fig b and Table).

A 13-year-old boy with stage 4 right adrenal neuroblastoma and multiple nodal, bone and bone marrow metastases underwent chemotherapy (HKPHOSG-NB-07 N7 protocol), gross total tumour resection, and autologous cord blood transplantation followed by immunotherapy. Complete remission was achieved but he had spinal relapse 4.5 years later presenting with cord compression at the level of T5-T9 vertebrae warranting emergency laminectomy and spinal tumour excision. He then received one cycle of temozolomide and irinotecan followed by adjuvant radiotherapy 30 Gy/10 Fr to T5-T9 vertebrae and a haploidentical transplant with maternal TCRαβ-depleted and CD45RA-depleted grafts. Total lymphoid irradiation at 8 Gy, fludarabine 150 mg/m², thiotepa 10 mg/kg and melphalan 140 mg/m² were prescribed as conditioning. Neutrophils and platelets were engrafted 10 and 11 days after transplantation, respectively, with 99% donor chimerism evident in marrow 30 days after transplantation. His post-transplant course was complicated by severe acute graft-versus-host disease (GVHD) involving skin (grades 2-3), liver (grade 2, biopsy-proven) and gut (grade 4) requiring prolonged and heavy immunosuppression including prednisolone, cyclosporine, and mycophenolate mofetil. He also had disseminated nocardiosis complicated by left lower lobe necrotising pneumonia, parapneumonic effusion, hydropneumothorax and bronchopleural fistula. Prolonged courses of antimicrobials were required and included meropenem, levofloxacin, ceftriaxone, cotrimoxazole, linezolid and amikacin. Due to liver GVHD, the patient had grossly deranged liver function 8 months post-transplantation with peak alanine aminotransferase level of 720 IU/L, aspartate aminotransferase level 506 IU/L, and gamma-glutamyltransferase level 1916 IU/L. His worst cholestasis occurred 2 years post-transplantation with total bilirubin level of 312 μmol/L and direct bilirubin level 270 μmol/L (Fig c and Table). He was noted to have a lipemic blood sample (Fig d) at this time and lipid profile revealed total cholesterol level of 13.6 mmol/L, LDL-C (calculated) level 11.4 mmol/L, HDL-C level 0.2 mmol/L and TG level 4.2 mmol/L (Table). Physical examination did not show any xanthoma. With the clinical context of severe cholestasis, LDL-C was measured directly and revealed discordance between the measured and the calculated value (1.7 mmol/L vs 11.4 mmol/L). Low Apo A1 (0.52 g/L) and elevated Apo B (1.7 g/L) levels were revealed. Trace chylomicron band, Lp-X band and faint lipoprotein Y bands were detected on lipoprotein electrophoresis (Fig d).

Liver GVHD is a known complication of allogeneic haematopoietic stem cell transplantation. It is characterised by elevation of hepatic enzymes, cholestasis, severe hypercholesterolaemia and hypertriglyceridaemia (in excess of 1000 mg/dL). In contrast to drug-mediated hypercholesterolaemia (cyclosporine, sirolimus, mycophenolate and glucocorticoids) when cholesterol level is usually <7.8 mmol/L (300 mg/dL) and mediated by LDL-C, hypercholesterolaemia caused by liver GVHD is mediated by cholestasis. In general, there are two main sources that contribute to the liver cholesterol pool, namely de novo (endogenous) cholesterol that is mainly synthesised in the liver, and dietary cholesterol (exogenous). Liver is the primary site of cholesterol biosynthesis and storage. It is also the principal site of sterol elimination by converting cholesterol to bile acids and removing free cholesterol as neutral sterols via biliary excretion.1 2 In liver GVHD-related cholestasis, impaired bile flow results in accumulation of cholesterol and bile salts, and hence elevated LDL-C level. Lipoprotein-X is another major cause of hyperlipidaemia in cholestasis when bile constituents reflux from the bile ducts or hepatocytes to the blood stream. Lipoprotein-X particles are formed when bile lipoprotein enters the blood stream and incorporates TG, Apo C and esterified cholesterol. Unlike LDL-C, Lp-X does not contain Apo B, the most important ligand to the hepatic LDL-C receptor. Therefore, Lp-X cannot be internalised into hepatocytes. Since Lp-X hypercholesterolaemia is not due to overproduction by hepatocytes, use of medications such as statins to downregulate cholesterol synthesis is ineffective.3 In addition, since Lp-X does not contain Apo B, which is the major component of LDL and one of the most important factors in the pathogenesis of atherosclerotic plaques, it is not atherogenic.4 Neither of our cases reported here had any complications of hypercholesterolaemia including exanthemata, retinal thromboembolism and pulmonary cholesteroloma. Nonetheless Lp-X may be associated with hyperviscosity syndrome and plasma exchange or apheresis may be indicated.5

As reported in the literature, hypercholesterolaemia secondary to intrahepatic cholestasis caused by liver GVHD can appear at any time between 2 months and 2 years after haematopoietic stem cell transplantation. The condition can be easily diagnosed by demonstrating discordance between calculated and directly measured LDL-C level, as well as lipoprotein electrophoresis. With the resolution of cholestasis, Lp-X will resolve with no specific treatment.

To conclude, severe hypercholesterolaemia mediated by Lp-X in post-haematopoietic stem cell transplantation patients with liver GVHD is a recognised yet overlooked phenomenon. Reports in the literature are limited for both adult and paediatric populations. To the best of our knowledge, this is the first report in Chinese children. Transplant physicians and endocrinologists should have an increased awareness of this association and avoid unnecessary and ineffective use of statins.

Concept or design: WYK Chan, JYL Tung.
Acquisition of data: WYK Chan, ECY Law, TK Ling, FCK Wong, JYL Tung.
Analysis or interpretation of data: All authors.
Drafting of the manuscript: WYK Chan.
Critical revision of the manuscript for important intellectual content: All authors.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Patients were treated in accordance with the Declaration of Helsinki. Consents for treatment, procedures and publication were obtained.

1. Fellin R, Manzato E. Lipoprotein-X fifty years after its original discovery. Nutr Metab Cardiovasc Dis 2019;29:4-8. Crossref

2. Nemes K, Åberg F, Gylling H, Isoniemi H. Cholesterol metabolism in cholestatic liver disease and liver transplantation: from molecular mechanisms to clinical implications. World J Hepatol 2016;8:924-32. Crossref

3. Zidan H, Lo S, Wiebe D, Talano J, Alemzadeh R. Severe hypercholesterolemia mediated by lipoprotein X in a pediatric patient with chronic graft-versus-host disease of the liver. Pediatr Blood Cancer 2008;50:1280-1. Crossref

4. Miida T, Hirayama S. Controversy over the atherogenicity of lipoprotein-X. Curr Opin Endocrinol Diabetes Obes 2019;26:117-23. Crossref

5. Heinl RE, Tennant HM, Ricketts JC, et al. Lipoprotein-X disease in the setting of severe cholestatic hepatobiliary autoimmune disease. J Clin Lipidol 2017;11:282-6. Crossref

Accelerating diagnosis and treatment of rifampicin-resistant/multidrug-resistant tuberculosis (MDR-TB) are key components of the World Health Organization’s End TB Strategy.1 Rifampicin resistance that arises from mutations outside the 81-base pair (bp) rifampicin-resistance determining region (RRDR) of the rpoB gene such as the I572F mutation nonetheless cannot be detected by existing World Health Organization–endorsed Xpert MTB/RIF assay or line probe assays (LPAs).2 Mutations located outside the rpoB hotspot may also be missed by the liquid medium–based BACTEC Mycobacteria Growth Indicator Tube (MGIT) culture system. We report a rare case of MDR-TB with rpoB I572F mutation (Escherichia coli numbering system) that was missed by LPA and liquid culture but confirmed by full rpoB gene sequencing to illustrate the negative impact of the mutation being undetected.

A 54-year-old Chinese man presented with symptoms of TB in July 2019. He had no history of anti-TB treatment. Chest radiograph on presentation showed bilateral cavitary lesions (Fig a). Sputum acid-fast bacilli smear examination was positive. Xpert MTB/RIF assay confirmed Mycobacterium tuberculosis (MTB)–positive/rifampicin-negative pulmonary TB. He was prescribed standard anti-TB treatment with isoniazid, rifampicin, ethambutol, and pyrazinamide. Initial sputum culture (performed with an MGIT culture system) later confirmed MTB organisms susceptible to all first-line drugs. Nonetheless lung shadows did not improve on serial chest radiographs (Fig b and c) despite good compliance with directly observed therapy. Sputum culture was transiently negative between November 2019 and January 2020. Culture result of a sputum specimen saved in October 2019 became available in January 2020 and showed MTB organisms resistant to isoniazid. The LPA (GenoType MTBDRplus Version 2.0; Hain Lifescience GmbH, Nehren, Germany) performed at that juncture showed an inhA C-15T mutation but no mutation in the rpoB gene. The regimen was switched to rifampicin, levofloxacin, ethambutol and pyrazinamide. Sputum culture reverted to positive in subsequent months and acid-fast bacilli smear also reverted to positive at 12 months. The LPA repeated at 12 months showed inhA C-15T and gyrA D94A mutations. Mutation of rpoB, rrs and enhanced intracellular survival genes was not detected although whole genome sequencing (WGS) performed on isolates obtained at 12 months revealed mutations of rpoB I572F, inhA C-15T, embB D354A, pncA D63G, and gyrA D94A. Whole genome sequencing performed retrospectively on isolates obtained earlier confirmed that rpoB I572F and inhA C-15T mutations had been present since the beginning of treatment. Culture result (using liquid culture system) of a sputum specimen saved at 12 months later became available and showed bacillary resistance to streptomycin, isoniazid, ethambutol and levofloxacin, but rifampicin resistance was again missed. In view of the gene sequencing result, the regimen was switched to a bedaquiline-containing regimen at 12 months. The patient responded well to treatment thereafter (Fig d).

Although most cases of rifampicin resistance are linked to mutations in the 81-bp hotspot region of the rpoB gene, notably mutations at codons 526 to 531, our case illustrates the occurrence of rare mutations outside the RRDR, namely I572F, and its negative impact on treatment outcome due to amplification of further drug resistance if left undetected. In our case, initial rifampicin resistance conferred by rpoB I572F mutation was missed by Xpert MTB/RIF assay, LPA and liquid medium–based culture system. Due to the exceptionally slow growth of this strain, prolonged incubation using MGIT was required and there was also difficulty in selecting log phase growth for subsequent drug susceptibility testing (DST). Not only was turnaround time increased, but initial isoniazid resistance was also missed although it was detected in the subsequent strains phenotypically. The patient therefore received an inadequate number of drugs during the early course of anti-TB treatment resulting in further acquired drug resistance to ethambutol, pyrazinamide, and levofloxacin. Amplification of drug resistance in our case likely emerged from the segregation of a single strain into two lineages of drug-susceptible and drug-resistant organisms under the selective pressure of insufficient TB therapy. This was suggested by the presence of both wild type and resistant subpopulations during the transition from susceptibility to resistance with regard to levofloxacin from the WGS data on D94A mutation.

Review of the database at the TB Supranational Reference Laboratory, Centre for Health Protection, Department of Health of Hong Kong revealed a total of five cases of MDR-TB with rpoB I572F mutation (including the present case) out of 340 rifampicin-resistant isolates between 2011 and 2020, corresponding to a prevalence of 1.5%. The prevalence of rpoB I572F in Hong Kong in this study is similar to that (2%) reported from a previous local study.3 On the contrary, a much higher prevalence of rpoB I572F mutation (corresponding to Ile491Phe mutation in MTB numbering system) has been reported recently in some countries with high TB prevalence such as Eswatini (formerly Swaziland) and South Africa (30% and 15%, respectively).4 The highly variable prevalence of rpoB I572F mutation in different geographical regions highlights the importance of expanding the geographical database of this mutation to better understand its global prevalence.

To improve the accuracy of phenotypic DST, the World Health Organization has recently lowered the critical concentration for rifampicin susceptibility testing in MGIT from 1 mg/L to 0.5 mg/L.5 The revised recommendation helps reduce but does not eliminate the discordance observed between phenotypic and molecular methods to detect rifampicin resistance and the potential false-susceptible results from phenotypic tests due to the presence of mutations outside the RRDR. Given the potential impact of rifampicin resistance conferred by an rpoB I572F mutation on treatment outcomes, and that an increasingly higher prevalence of such mutations has been reported recently in some countries, new molecular tests that expand the drug target coverage to help guide the formulation of treatment regimens are warranted. Expanding the target 81-bp hotspot RRDR of the rpoB gene to include codon 572 has been suggested.3 More recently, a multiplex allele-specific polymerase chain reaction (PCR) assay to detect I572F mutation in rpoB has been designed using a one-step real-time PCR.6 Although novel PCR assays may enable efficient and rapid detection of rpoB I572F mutation and are simpler than sequencing methods, their implementation requires further validation and strengthening of laboratory capacities.

Until novel PCR assays and WGS gain widespread use, clinicians should remain alert for the more rare rpoB mutations such as the I572F mutation, and communicate promptly with laboratories for further tests if a patient does not respond well to standard first-line treatment. This is vital even if Xpert MTB/RIF assay, LPAs or phenotypic DST do not suggest the presence of rifampicin resistance to ensure the patient receives an adequate number of effective drugs for treatment success. Our case also calls for continued surveillance of the prevalence of rpoB I572F mutation and other rifampicin-resistance conferring mutations outside the RRDR to inform region-specific TB diagnostic and treatment strategies.

All authors contributed to the concept or design of the study, acquisition of the data, analysis or interpretation of the data, drafting of the manuscript, and critical revision of the manuscript for important intellectual content.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

The authors have disclosed no conflicts of interest.

Part of the findings of this study has been presented at the Hong Kong Thoracic Society clinical meeting on 26 May 2022, which was an internal meeting attended by members of the Society held in virtual format.

This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

The patient was treated in accordance with the Declaration of Helsinki and provided written informed consent for publication.

1. World Health Organization. The end TB strategy. 2015. Available from: https://www.who.int/publications/i/item/WHO-HTM-TB-2015.19. Accessed 31 Jan 2023.

2. Nguyen TN, Anton-Le Berre V, Bañuls AL, Nguyen TV. Molecular diagnosis of drug-resistant tuberculosis; a literature review. Front Microbiol 2019;10:794. Crossref

3. Siu GK, Zhang Y, Lau TC, et al. Mutations outside the rifampicin resistance-determining region associated with rifampicin resistance in Mycobacterium tuberculosis. J Antimicrob Chemother 2011;66:730-3. Crossref

4. Variava E, Martinson N. Occult rifampicin-resistant tuberculosis: better assays are needed. Lancet Infect Dis 2018;18:1293-5. Crossref

5. World Health Organization. Technical report on critical concentrations for drug susceptibility testing of isoniazid and the rifamycins (rifampicin, rifabutin and rifapentine). Geneva: World Health Organization; 2021.

6. André E, Goeminne L, Colmant A, Beckert P, Niemann S, Delmee M. Novel rapid PCR for the detection of Ile491Phe rpoB mutation of Mycobacterium tuberculosis, a rifampicin-resistance-conferring mutation undetected by commercial assays. Clin Microbiol Infect 2017;23:267.e5-7. Crossref

In January 2017, a girl was born full term to non-consanguineous Chinese parents. Oesophageal atresia with a distal tracheo-oesophageal fistula was diagnosed soon after birth. Emergency surgical repair was attempted on day 1 of life but was complicated by complete transection of the left main bronchus that was subsequently repaired. The tracheo-oesophageal fistula was divided and a gastrostomy created. End-to-end anastomosis of the proximal and distal oesophagus was performed at age 6 months. Gastrostomy feeding was switched to jejunal feeding at 7 months because of gastroesophageal reflux and recurrent aspiration pneumonia. Attempted fundoplication at 17 months failed due to a small tubular stomach. She was prescribed oral ranitidine 30 mg three times daily.

At birth, the child’s haematology results were normal but by age 2 years she had developed persistent severe neutropenia (lowest neutrophil count 0.13 × 10⁹/L) and microcytic anaemia, with a lowest haemoglobin of 6.6 g/dL and lowest mean corpuscular volume of 65.5 fL. Platelet count was normal and she had no major infection or signs of anaemia. She was on continuous jejunal feeding with high-energy infant formula (Infatrini; Nutricia, Zoetermeer, The Netherlands) 640 mL daily as well as iron(III)-hydroxide polymaltose complex (IPC) 25 mg daily and multivitamin drops (Poly-Vi-Sol; Mead-Johnson Nutrition, Chicago [IL], United States) 1 mL daily. The high-energy infant formula provided 416 μg of copper daily, equivalent to 1.2 times the recommended dietary allowance.

Growth was satisfactory with her body weight at the 25th centile and height at the 10th centile. No new physical sign was detected. Peripheral blood smear showed occasional macro-ovalocytes. A dimorphic red cell picture was not seen and there were no hypersegmented neutrophils. Serum iron of 4.6 μmol/L (normal, 4-25 μmol/L), total iron-binding capacity of 103 μmol/L (normal, 41-77 μmol/L), and transferrin saturation of 4% (normal, 7%-44%) were suggestive of iron deficiency. Haemoglobin pattern analysis was negative for thalassaemia. Serum active vitamin B12 level was low at 23.6 pmol/L (normal, >46.2 pmol/L) and serum and red blood cell folate, serum bilirubin and lactate dehydrogenase levels were normal. Direct antiglobulin test, antinuclear antibody, C3, C4, anti-intrinsic factor antibody, antiparietal cell antibody and antineutrophil antibody were all negative.

Intramuscular vitamin B12 was given and ranitidine was stopped. The dose of IPC was increased to 15 mg twice daily and administered via the gastric tube.

Despite normalisation of serum iron and vitamin B12 level, anaemia and neutropenia persisted. Haemoglobin further dropped to 6.6 g/dL and red blood cell transfusion was required. She responded to a dose of granulocyte colony-stimulating factor with the neutrophil count rising from 0.25 × 10⁹/L to 1.74 × 10⁹/L. Nonetheless, her neutrophil count dropped to 0.33 × 10⁹/L 5 days later. Bone marrow aspiration and trephine biopsy was performed to evaluate the cause of refractory cytopenias and revealed reduced granulopoiesis, reactive histiocytosis and iron block. Vacuolated myeloid and erythroid precursors were observed (Fig 1). Megakaryopoiesis was adequate and no sideroblasts were evident on iron staining. These findings suggested copper deficiency or zinc toxicity. Serum copper was subsequently found to be <2.0 μmol/L (normal, 13-24 μmol/L), consistent with severe copper deficiency, whilst serum zinc level was normal.

Copper deficiency was treated with mineral mixture powder (Seravit; Nutricia) via the gastric tube from age 28 months as direct copper supplement was not available. The mineral mixture powder was administered at a dose of 2.5 g daily, providing 115 μg copper each day. On day 9 after commencement of copper supplementation, the absolute neutrophil count increased from the lowest level of 0.29 × 10⁹/L to 1.09 × 10⁹/L and haemoglobin level increased from the lowest level of 9.6 g/dL to 10.6 g/dL. The mineral mixture powder was further increased to 2.5 g twice daily. A small amount of milk was introduced to the stomach to allow absorption of the micronutrients. On day 16, the haemoglobin and neutrophil count normalised (haemoglobin, 12.2 g/dL; absolute neutrophil count, 2.03 × 10⁹/L) and by 7 weeks, serum copper had normalised. The mineral mixture powder was stopped after 8 weeks, at age 30 months. At age 35 months, haemoglobin, neutrophil count, copper, and iron levels remained normal (Fig 2).

The index patient presented with haematological abnormalities due to acquired copper deficiency following long-term jejunal feeding, which is not well reported in the literature.1 Jacobson et al2 reported three paediatric patients with exclusive jejunal feeding who developed cytopenias, one of whom had concurrent combined iron and vitamin B12 deficiency similar to our patient. Premature infants and children with intestinal failure on parenteral nutrition with inadequate copper supplementation are also at increased risk of acquired copper deficiency.3

Although the amount of iron, copper and vitamin B12 provided was above the recommended dietary requirement, deficiencies occurred because of problems in absorption. Most copper absorption occurs in the stomach and proximal duodenum. The acidic environment in the stomach facilitates solubilisation by dissociating it from copper-containing dietary macromolecules. Jejunal administration of IPC, multivitamin drops and infant formula bypassed the stomach and ranitidine reduced the acidity of the jejunal environment.

Copper-dependent enzymes are essential for normal function of the haematopoietic, skeletal, and central nervous systems. Although absent in the index patient, clinical signs of copper deficiency such as fragile, abnormally formed hair, depigmentation of the skin, oedema, myeloneuropathy, ataxia and cognitive deficits should be actively sought. Anaemia and neutropenia are the predominant haematological manifestations. Thrombocytopenia rarely occurs. Serum ferritin and erythropoietin are usually elevated. Serum ceruloplasmin is low. Anaemia is caused by the reduced activity of ceruloplasmin ferroxidase, copper/zinc superoxidase and cytochrome-c oxidase.4 Upon copper supplementation, neutropenia typically improves within a few weeks and anaemia improves within a few months. Cytoplasmic vacuolation in erythroid and myeloid precursors is the prominent feature in the bone marrow. Dysplastic features, such as megaloblastic changes and ring sideroblasts, may be observed.4 Vacuolated erythroblasts and myeloid precursors are classically observed in Pearson syndrome,5 acute alcoholism, chloramphenicol and linezolid toxicity, acute erythroid leukaemia and acute metabolic disturbance. These causes were unlikely in the index patient in the absence of an associated history or pathological features. Primary myelodysplasia is an important differential diagnosis but blood count recovery on copper replacement would not be expected.

This case highlights the importance of nutritional monitoring in patients receiving exclusive jejunal feeding. We recommend checking full blood count, liver and renal function tests, electrolytes, iron profile, vitamin B12, copper and zinc level every 3 months. Unexplained anaemia or neutropenia should prompt investigations for possible micronutrient deficiency to avoid unnecessary invasive investigations.

Concept or design: All authors.
Acquisition of data: WY Leung, CC So.
Analysis or interpretation of data: WY Leung.
Drafting of the manuscript: WY Leung, CC So.
Critical revision of the manuscript for important intellectual content: All authors.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

This case report received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

The patient was treated in accordance with the Declaration of Helsinki. The parent of the patient provided written consent for publication.

1. Barraclough H, Cooke K. Are patients fed directly into the jejunum at risk of copper deficiency? Arch Dis Child 2019;104:817-9. Crossref

2. Jacobson AE, Kahwash SB, Chawla A. Refractory cytopenias secondary to copper deficiency in children receiving exclusive jejunal nutrition. Pediatr Blood Cancer 2017;64:e26617. Crossref

3. Leite HP, Koch Nogueira PC, Uchoa KM, Carvalho de Camargo MF. Copper deficiency in children with intestinal failure: risk factors and influence on hematological cytopenias. JPEN J Parenter Enteral Nutr 2021;45:57-64. Crossref

4. Chen CC, Takeshima F, Miyazaki T, et al. Clinicopathological analysis of hematological disorders in tube-fed patients with copper deficiency. Intern Med 2007;46:839-44. Crossref

5. Knerr I, Metzler M, Niemeyer CM, et al. Hematologic features and clinical course of an infant with Pearson syndrome caused by a novel deletion of mitochondrial DNA. J Pediatr Hematol Oncol 2003;25:948-51. Crossref

Deep neck space infection is associated with significant morbidity and mortality, with the submandibular space being a common site of infection. Causes include odontogenic infections, submandibular sialadenitis, lymphadenitis, trauma, or surgery. This case report describes an uncommon cause of fish bone migration to the submandibular gland with consequent abscess formation. Apart from fish bones, a variety of foreign bodies in salivary glands has been described in previous case reports, including toothbrush bristles, slivers of fingernails, wood splinters, hairs or blades of grass.1

A 38-year-old woman with good past health accidentally ingested a fish bone in September 2020. She subsequently developed a sore throat, right neck pain and swelling. She attended the emergency department at Queen Mary Hospital 2 days later due to progressive symptoms. On admission, she had a fever of 38.1°C with blood pressure 161/80 mm Hg and pulse 113 beats per minute. Physical examination revealed a palpable tender right upper neck swelling. The right floor of mouth was oedematous and tender on palpation. At the opening of the Wharton’s duct, no foreign body was palpable, and no pus was evident. Flexible laryngoscopy of the pharynx and larynx was normal. An urgent computer tomography with contrast revealed a 14-mm linear foreign body at the right submandibular gland with surrounding right submandibular abscess (Fig 1).

Emergency exploration under general anaesthesia was performed. First, transoral exploration was performed through an incision at the right posterolateral floor of mouth to retrieve the foreign body and drain the pus. In view of negative transoral exploration, transcervical exploration was performed. An abscess cavity was revealed at the medial surface of the right submandibular gland and pus was drained. Right submandibular sialoadenectomy was performed and subsequently a 14-mm fish bone was found impacted at the medial surface of the right submandibular gland (Fig 2). A 15Fr silicon drain was inserted to the wound bed. The patient was monitored in the intensive care unit postoperatively and was extubated the day following surgery. Intravenous amoxicillin-clavulanate (1.2 grams every 8 hours) was prescribed. Culture of pus was negative. Fever and neck swelling subsided and the drain was removed on postoperative day 3. The patient was discharged on postoperative day 5 with diet well tolerated. At 2-week follow-up, the neck and oral wound were well healed. The hypoglossal nerve, lingual nerve and facial nerve function was intact.

A fish bone can usually be retrieved easily via endoscopy when lodged in the oral cavity, pharynx, or oesophagus. Nonetheless, some patients with a negative endoscopic finding may present later with neck swelling and fever that may indicate a deep neck space infection or migration of the foreign body to the neck. Extraluminal migration of foreign vein, subcutaneous neck, the thyroid gland, and the cervical spine has been reported in previous case reports.2

Migration of fish bone to the submandibular gland is uncommon. There have been two hypotheses for such an event: direct trauma and retrograde migration. Some theories suggest that continuous flow of saliva from ducts into the oral cavity, the duct orifice being mobile and able to twist in all directions, and small diameter of duct at the orifice render retrograde migration rare.1 In a previous study of sialoendoscopic assessment of patients with suspected obstruction in the ductal system of salivary glands, 3.9% had sialoliths related to fish bone, with the left submandibular gland being the dominant site (92.3%). Compared with the submandibular gland, foreign body at the parotid glands is much less common.3

Foreign body at the submandibular gland can have a variety of consequences. A patient with acute right submandibular sialadenitis due to an impacted fish bone has been reported who presented with tender submandibular swelling.4 The patient had no recall of foreign body ingestion. Plain radiograph of the neck revealed a radiopaque foreign body. The infection eventually required antibiotic treatment and submandibular gland excision, with identification of a fish bone at the excised gland. Another patient presented with chronic sialadenitis. A fish bone–induced sialolith was successfully removed with sialoendoscopy.5 A third patient presented with deep neck space infection with abscess formation, as in our patient.

Different techniques have been applied to remove foreign body from the submandibular gland.6 For sialoendoscopy, a higher success rate has been observed with foreign bodies in the distal duct while those located in the more proximal part and secondary branches of the Wharton’s duct have been difficult to remove. Surgery may be required if endoscopic treatment is not suitable, either using a transoral approach or transcervical approach with submandibular gland sialadenectomy. In this case report, a transcervical approach was eventually adopted.

The mainstay of treatment for submandibular abscess is airway protection, antibiotic treatment, and surgical drainage. Timely diagnosis and treatment are key to success. Despite being a rare cause, submandibular gland foreign body complicated by abscess formation should be considered as a differential diagnosis in patients with submandibular swelling who report swallowing of a fish bone.

Concept or design: RKY Tsang.
Acquisition of data: NHY Sun.
Analysis or interpretation of data: ATL Lau.
Drafting of the manuscript: ATL Lau.
Critical revision of the manuscript for important intellectual content: RKY Tsang.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

Dr Joseph CK Chung and Dr Sylvia SY Yu have made a substantial contribution in terms of academic advice to the publication of this case report.

This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

The study was conducted in accordance with guidelines of the Hong Kong West Cluster Research Ethics Committee (IRB Ref No.: UW21-094). Informed consent was obtained from the patient.

1. Su YX, Lao XM, Zheng GS, Liang, LZ, Huang XH, Liao GQ. Sialoendoscopic management of submandibular gland obstruction caused by intraglandular foreign body. Oral Surg Oral Med Oral Pathol Oral Radiol 2012;114:e17-21. Crossref

2. Hsu CL, Chen CW. A prolonged buried fish bone mimicking Ludwig angina. Am J Otolaryngol 2011;32:75-6. Crossref

3. Xie L, Zheng L, Yu C, et al. Foreign body induced sialolithiasis treated by sialoendoscopic intervention. J Craniofac Surg 2014;25:1372-5. Crossref

4. Riccio FJ, Scavo VJ. Unusual foreign body etiology of sialadenitis. Arch Otolaryngol 1967;86:210-2. Crossref

5. Iwai T, Sugiyama S, Hayashi Y, et al. Sialendoscopic removal of fish bone-induced sialoliths in the duct of the submandibular gland. Auris Nasus Larynx 2018;45:343-5. Crossref

6. Li P, Zhu H, Huang D. Detection of a metallic foreign body in the Wharton duct: a case report. Medicine (Baltimore) 2018;97:e12939. Crossref