Persistent eosinophilia following treatment for pediatric pyogenic liver abscess: A case report

Abstract

This case report describes a case of pyogenic liver abscess potentially associated with toxocariasis, a rare disease in the United States. A 3-year-old girl presented with new abdominal pain and fever. Contrasted computed tomography of the abdomen revealed a 2-cm hepatic abscess, which was subsequently drained. Culture of abscess fluid grew methicillin-resistant Staphylococcus aureus, and the patient was treated with intravenous vancomycin before transition to clindamycin and discharge home. Further laboratory evaluation yielded persistent eosinophilia and elevated IgE levels, with ambiguous primary immunodeficiency genetic testing. The patient’s hepatic abscess resolved; however, persistent eosinophilia prompted a parasitic workup, which returned positive for Toxocara IgG. The patient was thereafter treated with albendazole with subsequent decrease in eosinophil levels. This case report brings attention to a pediatric pathology previously unreported in the United States: toxocariasis potentially associated with development of pediatric methicillin-resistant S. aureus-associated pyogenic liver abscess.

Educational objectives

• Increase awareness of helminthic infection as a risk factor for pyogenic liver abscess, including in developed countries with lower incidence of helminthic infection.

• Identify approaches to the diagnostic workup in pediatric patients with pyogenic liver abscess, including relevant predisposing factors and their manifestation in laboratory evaluations.

Introduction

Pediatric pyogenic liver abscess (PLA) is rare in the United States, accounting for 13.5 per 100,000 pediatric hospitalizations, but has a nontrivial mortality rate of nearly 1%. ¹ There exist numerous predisposing factors to PLA, including appendicitis, cholecystitis, biliary atresia, and various immunodeficiencies. ² These immunodeficiencies include chronic granulomatous disease, autosomal dominant hyperimmunoglobulin E syndrome (AD-HIES), Wiskott-Aldrich syndrome, and Papillon-Lefevre syndrome.^3–5 The association of AD-HIES and Wiskott-Aldrich syndrome with both hepatic abscess and eosinophilia can prove a diagnostic challenge, as parasitic infection can also share these associations. However, because rates of helminthic infection are not as high in some regions of the United States as they are in more uniformly tropical countries, clinicians may pursue genetic testing for inherited conditions in PLA patients instead of potentially informative – or even diagnostic – evaluation for parasitic infection. Parasitic infection is both treatable and can cause long-term harm if left unmanaged; for example, untreated toxocariasis, which is associated with pyogenic abscess formation, ⁶ can result in visceral and ocular larva migrans. Still, no reports exist describing co-occurring toxocariasis and pediatric PLA formation in the United States. Therefore, we present a rare case in the United States of a 3-year-old girl presenting with abdominal pain and eosinophilia found to have a PLA in the setting of Toxocara infection.

Case

A 3-year-old female with past medical history of pica and two episodes of pneumonia presented to the emergency department (ED) with a chief complaint of 1 week of periumbilical abdominal pain, dysuria, sore throat, and fever with maximum temperature of 103.5°F. The patient had no nausea, vomiting, or changes in bowel habits. The patient was up to date on her vaccinations, and family history was notable for hepatitis A in her father and lupus in her maternal grandmother. The patient had exposures to dogs, cats, goats, rabbits, and a pet squirrel. She had no travel history outside of the southeastern United States. The patient was living with her siblings at home, none of whom had been sick, and she was not attending daycare.

Upon ED presentation, the patient’s pulse was 123 beats per minute, blood pressure was not recorded, respiratory rate was 24 breaths per minute, oxygen saturation was 99% in room air, and temperature was 100°F. On physical examination, the patient was somnolent and unresponsive to commands. Periumbilical abdominal tenderness and guarding were present. Skin examination was initially normal, with subsequent development of a pruritic, erythematous blanching rash in the left thigh, right lower abdomen, webs of the hands (sparing the palms), and right upper arm.

Laboratory evaluation on initial presentation revealed a negative rapid streptococcal test and a complete blood count with 15,840 per microliter white blood cells (normal range 4300–12,400) with a neutrophilic predominance of 69.3% and absolute neutrophil count of 10,980 per microliter (1120–8100), absolute eosinophil count of 2660 per microliter (0–300) with a proportion of 16.8%, hemoglobin of 10.6 g/dL (10.9–14.8), and platelets of 527,000 per microliter (150,000–450,000). Urinalysis was normal. Contrasted computed tomography (CT) of the abdomen and pelvis revealed a 2-cm hepatic abscess in segment IVa abutting the anterior abdominal wall, with possible rupture into the subcapsular space (Figure 1). Given these studies, the patient received one dose of 10 mg/kg metronidazole and one dose of 50 mg/kg ceftriaxone and was transferred to a tertiary care center.

Figure 1.

Axial (left) and sagittal (right) computed tomography views of hepatic abscess (red arrows) upon presentation.

At the tertiary medical center, further laboratory investigations revealed a C-reactive protein (CRP) of 164.6 mg/L (0.1–1.0), procalcitonin of 0.59 ng/mL (⩽0.24), normal gamma-glutamyl transpeptidase, and normal direct bilirubin. A complete metabolic panel demonstrated normal total protein level, hypoalbuminemia to 2.9 g/dL (3.8–4.7), aspartate aminotransferase level of 17 units/L (21–44), and alanine aminotransferase level of 9 units/L (9–25). Prothrombin time was elevated at 16.4 s (11.9–14.5), with a normal international normalized ratio (INR) of 1.3. Lipase was normal. Entamoeba histolytica IgG testing was negative. Blood cultures were collected and remained without growth throughout the patient’s hospitalization. Additionally, given the association between liver abscess and brain abscess ⁷ in the context of the patient’s altered mental status, a CT scan of the head with contrast was obtained and read as normal by a pediatric radiologist.

The patient was started empirically on intravenous metronidazole and ceftriaxone. The next day, the patient underwent ultrasound-guided percutaneous abscess drainage. Two milliliters of purulent fluid were recovered for diagnostic purposes, and post-procedure ultrasound demonstrated near complete resolution of the fluid collection. A culture of the abscess grew methicillin-resistant Staphylococcus aureus (MRSA) susceptible to clindamycin, doxycycline, trimethoprim-sulfamethoxazole, and vancomycin. The patient was transitioned to clindamycin (40 mg/kg/day intravenous given three times daily for 5 days) and vancomycin (15 mg/kg every 6 h for 3 days) after MRSA was isolated, with vancomycin discontinued once clindamycin susceptibility was documented. An echocardiogram showed no evidence of endocarditis. Based on the patient’s history of pneumonia and a new diagnosis of liver abscess, an immune workup was initiated. Serum IgA, IgM, and IgG were normal; IgE was elevated to 6660 IU/mL (0–50). Lymphocyte flow cytometry was significant for elevated proportions of CD3 and CD8 cells. The patient improved and was sent home with a 1-month course of oral clindamycin (40 mg/kg/day taken three times daily) and plans for further outpatient workup (see Figure 2).

Figure 2.

Clinical time course. Top panel: clinical events and medications from initial outpatient presentation to hospital discharge. Middle panel: IgE levels and absolute eosinophil levels from hospital discharge onward. Bottom panel: clinical events and medications from hospital discharge onward.

At 2-week follow-up after discharge (i.e., approximately 3 weeks after initial outpatient presentation), abdominal ultrasonography showed continued reduction in the size of the liver abscess from 2.1 × 2.4 × 2.3 cm (estimated volume 5.8 mL) to 1.6 cm in its longest dimension (estimated volume 0.8 mL) and with resolution of internal fluid. Improvement in CRP was also noted (7.6 mg/L from 70.9 mg/L). Eosinophilia persisted at 49.3% with an absolute eosinophil count of 8290 per microliter, and IgE remained elevated at 3857 IU/mL, although clinically, the patient was back to her baseline. Given the patient’s liver abscess, especially in the setting of elevated IgE and eosinophilia, testing for primary immunodeficiencies was initiated. Neutrophil oxidative burst test was normal. An Invitae Primary Immunodeficiency Panel was obtained and identified: (1) a heterozygous risk allele in NOD2 c.2104C>T (p.Arg702Trp); (2) a heterozygous pathogenic mutation in RMRP n_*25_*3dup (noncoding); and (3) a heterozygous variant of unknown significance (VUS) RMRP n.124C>A (RNA change). Specifically, the Arg702Trp mutation in NOD2 has been associated with a slightly increased risk of Crohn’s disease ⁸ and Yao syndrome ⁹ in heterozygotes, while the pathogenic mutation in RMRP is associated with cartilage-hair hypoplasia in homozygotes. ¹⁰ The VUS in RMRP has not yet been associated with any pathology. None of these identified variants were hypothesized by the care team to lead to a functional immunodeficiency and did not appear to correlate with her clinical presentation or susceptibility to Staphylococcal infections. No further intervention beyond trending IgE and eosinophil counts was indicated at that time. The patient’s eosinophilia persisted, with measurements taken at the 1st month after initial outpatient presentation (49.3% eosinophils, absolute count 8290 per microliter), one-and-a-half months after initial presentation (56.7% eosinophils, absolute count 9740 per microliter), three-and-a-half months after initial presentation (42.0% eosinophils, 5120 per microliter), and six-and-a-half months after initial presentation (46.8% eosinophils, 7900 per microliter). This prompted a broad workup for persistent eosinophilia, including hematology/oncology and allergy/immunology consultation, at approximately 7 months after initial presentation to rule out malignancy, infection, and immune dysfunction.

Tryptase, FISH PDGFRa/b panel, lactate dehydrogenase, uric acid, and chest X-ray were normal, as was serological testing for Strongyloides IgG and Histoplasma antibodies. Flow cytometry was repeated to evaluate T-cell populations and was interpreted as not suggestive of an underlying immunological process. Diphtheria IgG, tetanus IgG, and Streptococcus pneumoniae IgG 23 serotypes were all in protective range. IgG level was elevated at 1476 mg/dL (normal 500–1170), and IgA level was within normal limits at 116 mg/dL (normal 10–140). Enzyme-linked immunosorbent assay testing for Toxocara was positive (IgG = 71 units, normal <8), but no complications of toxocariasis beyond a potential predisposition toward PLA were noted. Based on guidelines from the American Academy of Pediatrics Red Book, ¹¹ treatment with albendazole 200 mg twice daily for 5 days was initiated and completed given that active, subclinical Toxocara infection could not be excluded. Eosinophilia and IgE levels initially improved; however, eosinophil count increased again (19.2% eosinophils, absolute count 2180 per microliter (30–510)) at follow-up approximately 3 months later (i.e., 10 months after initial presentation), so albendazole treatment (again with 200 mg twice daily for 5 days) was repeated with subsequent resolution of eosinophilia. While repeat testing 1 year later demonstrated a slight increase in eosinophil levels (6.8% eosinophils, absolute count 650 per microliter) and IgE (3163 IU/mL), this was deemed clinically insignificant given a lack of change in the patient’s clinical status, and the patient’s infection was considered sufficiently resolved.

In summary, the patient had eosinophilia and elevated IgE upon initial presentation with PLA. These both persisted after drainage and radiographic resolution of the abscess, prompting treatment with two courses of albendazole. Eosinophil counts temporarily dropped to normal levels and then stabilized to a nonclinically significant elevation with treatment; IgE remained elevated throughout the course of treatment. All treatment was well-tolerated and had no adverse effects according to the patient’s caregivers, who also ensured proper adherence.

Discussion

The lower incidence of parasitic infections in the United States compared with that in the developing world^12,13 may delay clinicians’ consideration of parasitic coinfection in patients with MRSA hepatic abscess. Indeed, most North American literature describing PLAs does not mention parasitic infection (including toxocariasis) as a potential risk factor for pyogenic abscess formation. PLAs are far less common in the United States than in more equatorial countries – the prevalence of PLA in the United States has been estimated at 22 per 100,000 hospital admissions versus rates as high as 1 case in 138 admissions in some parts of Brazil.^14,15 Furthermore, in contrast to Brazil, where PLAs have been found to affect mainly children, ⁶ over 90% of patients with PLA in the United States are adults, whose abscesses are overwhelmingly associated with structural biliary or hepatic pathology. ¹⁴ Given this, North American clinicians may be less likely to consider parasitic pathology as a predisposing factor to PLA formation.

Toxocara seroprevalence rates are estimated at 5.0% in the United States, ¹⁶ with higher rates in the southern regions of the country. ¹⁷ There is also epidemiological and experimental evidence that toxocariasis may predispose patients to PLA formation. In Brazil, the odds of Toxocara seropositivity were found to be significantly greater in PLA-affected patients than in controls. ⁶ Furthermore, one Brazilian case report of children with PLA found liver granulomas consistent in morphology with those produced by nematode larvae in nearly 30% of cases, far higher than the proportion of patients without PLA that had these types of granulomas (approximately 5%). The authors posited that Toxocara was the likely source of these granulomas. ¹⁸ Therefore, granuloma formation secondary to hepatic Toxocara translocation may serve as a predisposing factor for S. aureus colonization and PLA formation.

The immune response to helminthic infection itself may also predispose patients to PLA formation. Helminthic infection induces Th2-biased cellular responses and thereby increases levels of IL-4, IL-5, and IgE. This in turn may lead to a relative decrease in Th1 cells and therefore reduced antibacterial immune responses. ¹⁹ Th2-biased immune effectors can inhibit the Th1 response; indeed, clinical cases also exist of pediatric hepatic abscesses secondary to ascariasis in the setting of an exaggerated Th2/IgE response. ²⁰ By creating hepatic nidi for bacterial infection and impairing the immune system’s ability to respond to bacterial infection, helminthic infection can increase patients’ risk of developing pyogenic hepatic abscesses as well as pyogenic abscesses elsewhere.

Given that tissue biopsy was not obtained, the link between this patient’s toxocariasis and PLA remains speculative, and other factors beyond helminthic infection may have predisposed her to PLA. For example, the patient’s heterozygous Arg702Trp mutation in NOD2 has been associated with increased risk of clinically significant intra-abdominal bacterial infection in some patients post–liver transplant ²¹ ; however, the reported infections did not include liver abscess, and NOD2-associated immune dysfunction may only become clinically significant in the context of posttransplant immunosuppression, especially since the patient’s mutation has been found in healthy individuals in population-wide studies. ²² The patient’s otherwise negative genetic panel and flow cytometry ruled out other primary immunodeficiency syndromes, such as hyper-IgE syndrome, even given the patient’s persistently elevated IgE levels. Persistent elevation of IgE has been seen in other patients with Toxocara-associated liver abscess, even after treatment with albendazole. ²³ While none of the mutations found in the patient’s genetic panel for primary immunodeficiency (i.e., a heterozygous NOD2 risk allele for Crohn’s disease and Yao syndrome, a heterozygous pathogenic mutation in RMRP for cartilage-hair hypoplasia, and a separate heterozygous VUS in RMRP) were pathognomonic for an inherited immunodeficiency, it is still possible that these changes, together with other untested risk alleles, may confer the patient with a previously undescribed immunodeficiency. Taken together, the genetic testing results in isolation were unlikely to explain the patient’s clinical phenotype.

The virulence capacity of the patient’s strain of MRSA may have also predisposed her to PLA. While molecular testing was not performed on her MRSA isolate, MRSA strains positive for staphylococcal cassette chromosome mec type IV, Panton-Valentine leukocidin, and toxic shock syndrome toxin-1 are associated with severe MRSA infection in pediatric patients, even in those without underlying immunodeficiency. ²⁴

It is also worth noting that toxocariasis serology testing has inherent limitations. For one, IgG seropositivity may reflect a prior resolved infection rather than active infection. In addition, the sensitivity and specificity of laboratory testing can vary significantly based on the particular IgG heavy chain examined, with sensitivity rates ranging from 50% to 98%, and specificity rates from 71% to 81%. ²⁵ This issue is further exacerbated by the potential of serologic cross-reactivity with other helminthic infections, especially ascariasis, angiostrongyliasis, and trichinosis. ²⁵ Still, repeat testing or quantitative trends, while not necessarily diagnostic, may occasionally help with the clinical interpretation of these tests. Strongyloidiasis testing was performed in the patient to rule out cross-reactivity with this pathogen, but the possibility of a false-positive Toxocara serology in the setting of another helminthic infection, perhaps ascariasis, cannot be excluded.

Toxocara is most often transmitted to humans via ingestion of soil containing infectious eggs, often present in canine or feline feces, and can also be transmitted by the consumption of undercooked meat containing encapsulated larvae. ²⁶ We hypothesize that the patient’s exposure to dogs and cats at home – especially in the setting of pica – predisposed the patient to Toxocara infection, which in turn may have predisposed her to PLA formation in the setting of toxocariasis visceral migrans. Pediatric PLA formation in the context of toxocariasis illustrates the necessity of a broad differential diagnosis when considering the contributors to PLA.

Conclusion

PLA is a rare but serious pediatric condition. Multiple pathologies may contribute to the development of PLA and include intrabdominal infection (e.g., appendicitis), inherited immunodeficiencies, and helminthic infection. Some heritable immune conditions share similar laboratory manifestations with parasitic infection, including elevated IgE levels and immunodeficiencies; however, due to the relatively low prevalence of helminthic infection in nontropical areas, clinicians may be more likely to pursue immune, rather than parasitic, workup for PLA patients. The current case report presents a North American pediatric patient who presented with PLA and persistent eosinophilia after its treatment, and whose only identifiable potential predisposing factor for PLA was Toxocara infection. The Toxocara infection, while not definitively confirmed to contribute to PLA susceptibility, may have facilitated PLA development. Therefore, when treating pediatric patients with PLA, clinicians should keep in mind that parasitic infection – along with inherited immune conditions – may predispose patients toward this condition.