We describe 9 cases of severe, life-threatening liver abscesses in X-linked CGD that progressed despite appropriate antibacterial therapy and drainage, where appropriate, that responded to moderately high doses of corticosteroids tapered over several months. All patients presented with nonspecific symptoms including fever, malaise, abdominal pain, and elevated inflammatory markers; 6 had prior histories of liver abscess (). This agrees with previous data that prior liver abscesses predict recurrence [12
]. The best results were obtained when the steroids were initiated at a moderately high dose and their administration was prolonged; local hepatic inflammation often flared when steroids were tapered. In those patients in whom extended steroids were used along with appropriate antibiotics, no disease worsened, and 8 of 9 patients achieved clinical cure without surgery.
Hepatic abscesses are a common complication of CGD, are typically due to S. aureus
, and are difficult to treat [14
]. Lublin et al reported that 16 of 61 cases required multiple procedures with an overall surgical complication rate of 56% [3
]. Despite the low mortality rates of liver abscesses treated with antibiotics and surgery, the morbidity is high and there is a high rate of recurrence. Perhaps more important, hepatic dysfunction with portal venopathy and nodular regenerative hyperplasia (NRH) are highly associated with liver abscesses and are major predictors of mortality in CGD [12
], independent of level of superoxide production [1
]. The mechanism is presumed to be noncirrhotic portal hypertension caused by microvascular insults from repeated abscesses [12
]. It is possible that hepatic regeneration after surgery causes venopathy and NRH. However, it remains unclear whether liver abscesses are causes, effects, or simply parallel markers of this predisposition to severe liver dysfunction.
Dysregulated inflammation in CGD is probably important in the late complications of the disease. Morgenstern et al demonstrated that both early and delayed inflammation was exaggerated in X-CGD mice independent of infection [15
]. Proposed mechanisms of hyperinflammation include defective neutrophil apoptosis [16
]; skewed nuclear factor-κB signaling [17
]; upregulation of tumor necrosis factor α, interleukin (IL) 17, IL-6, and granulocyte colony-stimulating factor [18
]; impaired degradation of leukotriene B4
and C5a [19
]; prolonged IL-8 messenger RNA activation [20
]; impaired activation of Nrf2 [21
]; defective tryptophan catabolism in mice [22
] but not humans with CGD [23
]; and deficient peroxisome proliferator-activated receptor γ activation [24
]. Regardless of the specific mechanisms of hyperinflammation, corticosteroids reduce the activation, proliferation, and differentiation of many cells involved in inflammation, including macrophages and lymphocytes, the key components in granuloma formation [25
There are precedents for the successful use of steroids in CGD during acute infection when coadministered with appropriate antimicrobials. Mulch pneumonitis appeared to benefit from steroid therapy when administered along with adequate antifungals [11
]. The clinical and radiographic pattern of CGD mulch pneumonitis patients is similar to that seen in hypersensitivity pneumonitis, in which exuberant host response to environmental pathogens is typically treated by steroids [26
]. Similar results have been reported in patients with invasive pulmonary Nocardia
], chronic cystitis [28
], and in 2 cases of liver abscesses [10
]. The overlap between this aspect of CGD and hypersensitivity reactions is provocative.
The patients in this report had either inoperable disease, were too ill to undergo hepatic resection, or refused surgery. Corticosteroids as adjuncts to antibiotics were able to overcome weeks to months of static or progressive liver abscesses in patients with CGD. Corticosteroids were started at a dose of approximately 1 mg/kg daily and continued for an average of 2–3 weeks before being tapered over several months (). Within days of initiating corticosteroid therapy, patients exhibited clinical improvement, substantial decrease in lesions on imaging, and decrease in inflammatory markers. There was a fast decline in CRP and ESR after initiation of corticosteroids that continued to slowly decline over the ensuing months. All patients were continued on organism-specific antimicrobials, and only 1 of 9 eventually required surgery. Interestingly, the patient who required surgery (patient 3) had received the least and shortest course of steroids. Although there was no uniform steroid taper, the average duration was 3 months. Faster taper was associated with symptom, sign, and radiographic exacerbation, resulting in addition of antimicrobials and increased corticosteroids. Complications of steroid therapy were transient sleep disturbances, typical hyperphagia, and mild weight gain, all of which were reversible.
As is often true in retrospective case series of patients managed by independent groups, the treatments, doses, and durations of therapy received by the patients varied. However, by pulling together 9 cases of liver abscess in CGD with somewhat different outcomes, we can infer some general features, recognizing that prospective trials of surgery versus steroid therapy are unlikely to be performed. The duration of steroid therapy appears to be important, with radiographic evidence of exacerbation as doses became very low. In our series, the only patient who required open surgical excision after treatment with corticosteroids was not the one with the largest abscess, but the one who received the least steroids. In contrast, long-term high-dose steroids have very real consequences, including other opportunistic infections and metabolic effects. The necessity for intravenous antibiotic therapy is also unclear, but the coadministration of antibiotics with steroids is sensible and probably required.
Liver abscesses in CGD have not been eliminated by antibiotic prophylaxis, are generally survivable, and often recur. However, adult-onset liver dysfunction is highly associated with adverse outcomes in CGD [12
]. Interestingly, the mortality curves between high-risk and low-risk forms of CGD do not diverge until adulthood, implying that accumulated toxicities of the disease may play a role in outcome [1
]. We cannot know at this point whether the long-term consequences of steroid management of CGD liver abscesses are preferable to surgical management. However, it is clear that in the short term, at least in some cases, steroids coadministered with antibiotics allow cure of CGD liver abscesses without surgery.