Sinonasal mucormycosis and liposomal amphotericin B: A quest for dose optimization.

OBJECTIVES: Sinonasal mucormycosis is a serious fungal infection. Early diagnosis and prompt antifungal therapy along with surgical intervention is the key to its management. Liposomal amphotericin B (LAmB) given intravenously is the antifungal agent of choice. However, the current literature is not clear on its optimum dosage. We did a retrospective study to find the optimum dose of LAmB in cases with sinonasal mucormycosis.
MATERIALS AND METHODS: Thirty patients diagnosed with mucormycosis involving sinonasal, rhino-orbital, or rhino-orbito-cerebral regions and receiving only LAmB as pharmacotherapy were included in our retrospective study from 2017 to 2020. A multiple logistic regression model was developed to correlate the total dose of LAmB and other parameters with the final outcome which was defined clinico-radiologically as improved, worsened, or death. The dose of LAmB which led to the first significant change in urea, creatinine, and potassium levels was also determined.
RESULTS: The model showed a good fit in goodness-to-fit analysis (Pearson = 0.999, deviance = 0.995), while the likelihood ratio was statistically significant (0.001). The overall model prediction was 83.3%. However, the correlation of outcome with any of the variables, including mean LAmB dose per kilogram (82.2 ± 13.02 mg/kg), was statistically not significant.
CONCLUSION: Many patient-related factors (such as age, comorbidities, extent of the disease, and side effects from LAmB therapy), which vary on a case-to-case basis, contribute to the outcome in a mucormycosis patient. The optimum dose of LAmB for improved outcome still requires individualization guided by experience, till well-designed studies address the question.

Introduction

Mucormycosis is a life-threatening infection caused by a variety of fungal species in the order Mucorales, most common being Rhizopus, Lichtheimia, and Mucor.[12] In tropical regions like India, Apophysomyces elegans is also commonly seen.[23] It is an invasive fungal infection (IFI) that usually occurs in patients who are immunocompromised, those with diabetes mellitus, neutropenic conditions, hematological malignancies, solid organ or stem cell transplantation, or skin-penetrating injuries.[456] Clinically, it can involve different regions which include rhino-orbito-cerebral, cutaneous, pulmonary, gastrointestinal, and disseminated (involving 2 or more regions). Rarely, it may be isolated from a single organ as well (renal/lymph node/parotid/ear/heart).[27]

Early diagnosis is the key to treatment and management approaches including prompt antifungal therapy, surgical debridement, and addressing the underlying predisposing condition. In most situations, intravenous amphotericin B is used as the first line of antifungal treatment, with the liposomal form being less nephrotoxic than the conventional form, and thus able to be given for longer periods.

For most patients, 5–7.5 mg/kg/day of liposomal Amphotericin B (LAmB) is considered reasonable for most patients. In case of central nervous system involvement, a higher dose of 10 mg/kg/day is recommended.[5] Based on literature and international consensus, the European Organization for Research and Treatment of Cancer (EORTC)/Mycosis Study Group (MSG) proposed research-oriented definitions for the IFI and categorized them as “proven,” “probable,” and “possible.”[8] However, a clear guideline on the dose and duration of LAmB therapy is lacking. The dose varies on a case-to-case basis depending on factors such as comorbidities, patient age, extent of disease, need for surgical intervention, and treating physician's experience. Therefore, the recommended dose of 5–7 mg/kg/day may not result in improved outcomes for all cases.

We retrospectively reviewed data of patients with mucormycosis involving sinonasal/rhino-orbital/rhino-orbito-cerebral region, who presented at our tertiary care institution, and aimed at finding the most appropriate dosage of LAmB in such cases.

Materials and Methods

This was a retrospective study done in the department of otolaryngology and pharmacology at an apex tertiary care institution. The hospital medical records were perused after clearance from the Institute Ethics committee (IEC-59/8.1.21, RP-07/2021). The first and last patients included were hospitalized in June 2017 and November 2020, respectively.

Patients diagnosed with “Proven” and “Probable” invasive mucormycosis infection (as per the EORTC/MSG definition) involving sinonasal/rhino-orbital/rhino-orbito-cerebral regions and receiving only LAmB as pharmacotherapy were included. The patients receiving any other antifungal drug in addition to LAmB were excluded [Figure 1]. The microbiological diagnosis was established either by examination of 10% potassium hydroxide mount of nasal/palatal crust/tissue biopsy specimen or by a histopathological diagnosis of tissue. None of the patients enrolled in our study had laboratory-diagnosed coronavirus disease of 2019 (COVID-19).

Figure 1

Patient enrolment and outcome

The final outcome was categorized into the following three categories:

Improved: Both clinical and radiological improvement

Clinical improvement: No clinical evidence of residual disease (necrotic tissue or bone/residual fungal debris/osteomyelitic bone)

Radiological improvement: No evidence of residual disease on computed tomography (CT) in all cases or magnetic resonance imaging (MRI) in cases involving orbital apex or those with intracranial extension.

Worsened: Either clinical or radiological deterioration

Clinical deterioration: Progression of disease (involvement of new regions/progression of necrosis) or no clinical improvement despite giving LAmB and/or possible surgical intervention

Radiological deterioration: Progression of disease (involvement of new regions) or persistent residual disease on CT or MRI despite giving LAmB and/or possible surgical intervention.

Death: Patients who expired during inhospital treatment.

The data were retrieved, cleaned, and entered into MS Excel. Analysis was done using IBM SPSS Statistics for Windows, Version 23.0. Armonk, NY: IBM Corp. A multiple logistic regression model was developed to correlate dependent variables such as duration of symptoms, comorbidities, extent of the disease, surgical intervention done, duration of hospital stay, and total dose of LAmB with the independent variable which was the final outcome of the patients (improved, worsened or death). Results were presented as mean ± standard deviation.

Results

The demographic details, clinical parameters evaluated, and mean total dose of LAmB in our study participants are given in Table 1. The baseline serum urea, creatinine, and electrolytes (sodium and potassium) levels were analyzed for all the patients. Table 2 shows mean of baseline levels and significant change in the levels of urea (>50% increase from baseline), creatinine (>50% increase from baseline), and potassium (<3.5 mmol/L), for the first-time, during treatment with a certain dose of LAmB. Raised serum urea and creatinine levels (>50% increase from baseline) were seen in 50% of patients. Out of these 15 patients, 9 had complete recovery and 4 had partial recovery (see discussion). Two patients with preexisting chronic kidney disease had raised baseline serum urea and creatinine levels. Hypokalemia (K + levels <3.5 mmol/L) was seen in 80% of the patients.

Table 1

Demography and clinical parameters (n=30)

Parameter	Number of patients
Mean age (years), mean±SD	44.3±15.3
Gender (%) (male: female)	66.6:33.3
Site of infection, n (%)
Unilateral disease	25/30 (83.3)
Sinonasal	8/25 (32)
Sino-orbital	12/25 (48)
With intracranial but extradural extension	1/25 (4)
With intracranial and intradural extension	4/25 (16)
Bilateral disease	5/30 (16.6)
Right-sided sinonasal and left-sided sino-orbital	1/5 (20)
With intracranial but extradural extension	2/5 (40)
With intracranial and intradural extension	2/5 (40)
Mean duration of symptoms (days), mean±SD	34.6±32.8
Comorbidities, n (%)
No comorbidity	5/30 (16.6)
1 comorbidity	12/30 (40)
DM	11/12 (91.6)
Postrenal transplant	1/12 (8.3)
>1 comorbidity	13/30 (43.3)
DM + HTN	5/13 (38.4)
DM + HCV	2/13 (15.3)
DM + Treated pulmonary tuberculosis	2/13 (15.3)
DM + CKD	1/13 (7.6)
DM + CKD+Chronic liver disease	1/13 (7.6)
DM + Coronary artery disease	1/13 (7.6)
DM + HTN+HCV	1/13 (7.6)
Number of surgeries, n (%)
No surgery	1/30 (3.3)
Single procedure	22/30 (73.3)
Endoscopic debridement	13/22 (59)
External debridement	9/22 (40.9)
Multiple surgical debridement procedures	7/30 (23.3)
Mean number of days admitted in hospital (days), mean±SD	38.5±27.8
Mean total dose of intravenous amphotericin B
Expressed as gram, mean±SD	5.29±1.93
Expressed as milligram/kilogram, mean±SD	82.19±37.0
Expressed as milligram/kilogram/day	2.13
Final outcome, n (%)
Improved	15/30 (50)
Worsened	6/30 (20)
Expired	9/30 (30)

SD=Standard deviation, DM=Diabetes mellitus, HTN=Hypertension, HCV=Hepatitis C virus, CKD=Chronic kidney disease

Table 2

Baseline level and first significantly changed level of Urea, Creatinine and Potassium after administration of LAmB

Parameter	Baseline level (mean ± SD)	First significantly changed level (mean ± SD)	Dose (mg) of LAmB given (mean ± SD)
Urea (mg/dl)	31.5±8.40	49.5±7.28*	1152.5±910.14
Creatinine (mg/dl)	0.8±0.36	1.6±0.29*	1612.5±700.18
Potassium (millimoles/l)	3.9±0.28	3.02±0.28**	1703.1±1280.18

SD=Standard deviation, * >50% increase from the baseline,

** <3.5 millimoles/l

The model showed a good fit in goodness-to-fit analysis (Pearson = 0.999, deviance = 0.995), as shown in Table 3, while the likelihood ratio was statistically significant (0.001). The overall model prediction was 83.3%. The prediction for improved and expired outcomes was 100%, while for the worsened was 16.7%. The correlation of outcome with any of the variables was not statistically significant [Table 4]. The model also did not show any correlation between the mean LAmB dose per kilogram with the outcome. Table 5 shows the dose and duration of LAmB therapy.

Table 3

Goodness of fit of multiple logistic regression model

	χ 2	df	Significance
Pearson	20.281	44	0.999
Deviance	23.735	44	0.995

Table 4

Parameter estimates of multiple logistic regression model

Outcome*	Coefficient	Significance	95% CI for Exp (B)
			Lower bound	Upper bound
Worsened intercept	−1.390	0.554
Age	0.033	0.533	0.933	1.145
Sex	−0.938	0.466	0.031	4.874
Site	0.242	0.720	0.338	4.805
Duration of symptoms	0.003	0.836	0.974	1.033
Number of surgeries	−0.250	0.812	0.099	6.102
Number of days admitted	−0.002	0.961	0.934	1.068
Dose per kilogram	−0.010	0.777	0.926	1.059
Expired intercept	−314.559	0.999
Age	6.743	0.996	0.000	**
Sex	242.458		1.987E+105	1.987E+105
Site	130.079	0.997	0.000	**
Duration of symptoms	−0.791	1.000	0.000	**
Number of surgeries	81.759	0.999	0.000	**
Number of days admitted	−6.098	0.998	0.000	**
Dose per kilogram	−3.803	0.995	0.000	**

*Reference category=Improved, **Floating-point overflow occurred while computing, and the value is set to system missing. CI=Confidence interval

Table 5

Dose and duration of liposomal amphotericin B expressed as mean±standard deviation with reference to the outcome

Outcome	Number of patients	Total dose (g)	Duration of LAmB therapy (days)
Improved	15	5.72±2.10	46.4±32.3
Worsened	6	6.13±1.26	41±16.8
Died	9	4.04±1.48	23.7±20.9

LAmB=Liposomal amphotericin B

Discussion

Mucormycosis is a life-threatening condition that can progress rapidly.[6] A comprehensive systemic review and meta-analysis of case reports of mucormycosis by Jeong et al. concluded amphotericin to be the mainstay antifungal agent for pharmacotherapy of mucormycosis.[7] Lipid-based amphotericin B has less renal toxicity and can be given for a longer duration of time.[57] However, international consensus for dose and duration of LAmB is lacking. A prospective study in 40 patients used a relatively higher dose of LAmB (10 mg/kg/day) for the treatment of mucormycosis in the initial stages and showed no change in mortality at high dose during the first 7 days of treatment.[6] According to the AmBiLoad trial in IFI, 10 mg/kg/day dose of LAmB did not improve efficacy but increased the toxicity and cost of treatment as compared to 3 mg/kg/day.[9] This trial, however, included only three and two patients, in the respective treatment arms, who had IFI due to Zygomycetes. Most of the patients having IFI were due to Aspergillus (91 in 10 mg/kg/day arm and 103 in 3 mg/kg/day arm). There was a statistically significant difference in toxicity between the two arms – nephrotoxicity and hypokalemia being higher in 10 mg/kg/day arm.

In our study, the duration and dose of LAmB varied from case to case. A high SD of the LAmB, almost approaching the mean, suggests wide variation in the dose of LAmB administered on a case-to-case basis. The observation re-emphasizes the fact that the dose and duration of amphotericin for mucor have not been standardized or systematically studied. We further tried to estimate the total dose of LAmB required for clinical and/or radiological improvement. However, we did not find any statistically significant contributor to the outcome, and the total dose of LAmB administered did not have any correlation with the outcome. In addition, the mean total dose administered in our patients was 2.1 mg/kg/day which was lower than the general recommendation of 5–7 mg/kg/day. Our data suggest a cautious approach while deciding the daily dose of LAmB to prevent life-threatening hypokalemia or nephrotoxicity. As discussed later, none of our patients required dialysis or developed life-threatening hypokalemia.

In India, most cases of mucormycosis are seen during monsoon and autumn.[2] In our study, 70% of the patients presented in the months of July to November. A tertiary care center in India shared its experience with Zygomycosis over 10 years and reported that rhino-orbito-cerebral mucormycosis was the most common type (44.2%) among 129 patients. Analysis of the outcome of the therapy was done in 33 patients (24 rhino-orbito-cerebral cases). Of these, 15 patients received amphotericin B alone (total dose 3–4 g) and 16 were treated by surgical debridement along with amphotericin B (dose not specified). Amphotericin-B given along with aggressive debridement was found to work best in more than 80% of patients improved with the combination.[10]

The angiotropism of Mucorales species causes blood vessel thrombosis and finally necrosis, resulting in low drug penetration.[56] Therefore, surgical removal of the devitalized tissue, whenever possible, forms an important part of the management of mucormycosis and has improved outcomes when compared to those getting only antifungal therapy.[145] In our study, all patients except one underwent surgical procedures in addition to LAmB therapy. Therefore, there was no group to compare the effect of surgical intervention on the outcome. The one patient who did not undergo surgery had multiple comorbidities (diabetes mellitus, chronic kidney disease, and chronic liver disease) and was not fit to undergo surgery under general anesthesia. The number of surgical debridement or number of surgical procedures performed had no correlation with the extent of the disease or outcome. Every patient was assessed clinically on daily basis and if there was suspicion of significant residual disease after first surgery or recurrence of disease (evident either clinically or radiologically), then he/she was taken for revision surgery or debridement.

LAmB has significantly less renal toxicity as compared to conventional amphotericin B (amphotericin B deoxycholate), as the drug gets locked in the liposomes and the particles have small size and negative charge.[1112] Although LAmB offers less toxicity than conventional amphotericin B, the reduction in adverse events, particularly nephrotoxicity, comes at a significantly higher cost.[1314] Such formulation may not be affordable in large quantities at the community level. Recently, there was an epidemic of COVID-19-associated mucormycosis in India.[1516] Many hospitals had to revert to conventional amphotericin B during this time due to high demand, short supply, and relatively higher cost of LAmB. The effect of COVID-19 or steroids used during the pandemic on dose and duration of LAmB or conventional amphotericin B are not known, as there is insufficient literature focusing on the difference in dosage of amphotericin used in COVID-19-associated mucormycosis versus non-COVID-19 mucormycosis. What is known is that there are less nephrotoxicity and fewer infusion-related reactions with continuous infusion of conventional amphotericin B over 24 h as compared to four-hourly to six-hourly infusions.[171819] Conventional amphotericin B can also be admixed with locally prepared lipid emulsions. Although there is no standardized protocol for making such extemporized mixture, a meta-analysis has shown its efficacy to be similar to LAmB and nephrotoxicity lower than conventional amphotericin B. This meta-analysis also compared LAmB with conventional amphotericin B and concluded that liposomes reduce nephrotoxicity of amphotericin by 18.1% ([99% confidence interval, 0.36–0.64]; n = 1233).[20]

The dose of LAmB at which renal toxicity starts is unclear, but it has been shown in animal studies that even at a higher dose range of 10–15 mg/kg/day, LAmB causes less toxicity as opposed to conventional amphotericin.[212223] LAmB has been found to have a minimal effect on renal function of critically ill patients who had elevated baseline serum creatinine, suggesting that LAmB can be used in such patients, independent of the renal function status at the time of treatment initiation.[24] In our study, 50% of the patients developed acute kidney injury which was defined as >50% increase in serum creatinine levels in accordance with the previously published studies.[2526] In these patients, LAmB therapy was halted for 24 h and 0.9% sodium chloride was given for salt loading as a toxicity-prevention strategy described widely.[27282930] LAmB therapy was resumed after 24 h, at a reduced dose than the previous day, after re-checking the creatinine levels. Four patients did not recover completely, but nine patients had complete recovery which was defined as a return to within 10% of pretreatment serum creatinine. None of the patients required dialysis. In a study to assess renal recovery following LAmB nephrotoxicity, the results suggested that the dose of LAmB dose had no impact on the probability of renal recovery.[31] Currently, salt loading by infusion of normal saline, avoidance of dehydration, or concomitant use of other nephrotoxic drugs appears to be a safe and effective strategy to prevent LAmB-induced nephrotoxicity. N-acetyl cysteine has been tried for prophylaxis in animal studies, but the evidence is inconclusive in humans.[323334]

Hypokalemia (K + levels <3.5 mmol/L) was seen in 80% of the patients in our study, but none of the patient (including those who worsened or expired) developed life-threatening hypokalemia. It has been shown in the studies that early supplementation of potassium, within 2 days of starting LAmB, is an independent factor for reducing the risk of LAmB induced hypokalemia and it is necessary to start potassium supplementation before the levels drop to <2.83 mEq/l.[3536] We also started our patients on either intravenous or regular oral potassium chloride syrup (20–40 mEq/day) after detection of first significantly lower potassium value (<3.5 mmol/L) and dose of LAmB therapy was adjusted daily to maintain potassium levels above 3.5 mmol/L. No other adverse effects such as allergic reaction or anaphylaxis were seen in our patients. The dose of LAmB was altered daily in 80% of the patients based on daily urea, creatinine, and potassium levels. The amphotericin blood levels were not checked in any of the patient.

Although the definitive dose and duration of LAmB for expecting an improved outcome in sinonasal mucormycosis remain debatable, its dose in visceral leishmaniasis or Kala-azar is well defined. It is currently recommended as a first-line antileishmanial drug.[37] The recommended dosage regimen is intravenous administration of 3 mg/kg/d LAmB on days 1–5, 14, and 21.[38] The regimen accounts for a total dose of 21 mg/kg and 20 mg/kg has been shown to have a 99% initial cure and low relapse rates.[39] The National Kala-Azar Elimination Programme of India uses 10 mg/kg LAmB as a single dose which also has a more than 95% cure rate.[40] Conventional amphotericin B also has similar efficacy to LAmB and can be used as a short-course regimen.[4142]

Our study focussed on sinonasal mucormycosis, but we included patients with orbital and intracranial extensions as these were extension of the infection from sinonasal region only. The study was limited by small sample size and retrospective review. Large-scale randomized controlled trials are needed to define the optimum dose which can lead to a decrease in mortality of patients with invasive fungal disease.

Conclusion

Sinonasal mucormycosis is a disease with a high mortality rate. Amphotericin B is an established pharmacotherapy, and its liposomal form has less toxicity than the conventional form. Surgery remains the cornerstone of management in such patients in addition to pharmacotherapy. The optimum dose of LAmB required for complete response in invasive fungal disease is not well established. There are many patient-related factors (such as age, comorbidities, extent of the disease, and side effects from antifungal therapy) which contribute to the response of the patient and should be considered in estimating the total dose of LAmB therapy. However, all these factors vary on a case-to-case basis and our study did not find any significant correlation between any of these factors and final outcome of the patient. Hence, the optimum dose of LAmB required for the improved outcome should be estimated on an individual basis.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.