COVID-19-associated rhino-orbital-cerebral mucormycosis: A systematic review, meta-analysis, and meta-regression analysis.

BACKGROUND: Till now, no meta-analysis is available to address the clinical profile, risk factors, different interventions, and outcomes among COVID-19-associated rhino-orbito-cerebral mucormycosis (C-ROCM) cases.
MATERIALS AND METHODS: Eight literature databases were screened using appropriate keywords from November 1, 2019, to June 30, 2021. The objectives were to analyze the clinical and microbiological profile, risk factor/comorbidity, intervention, and outcome. "R-metafor package" was used for analysis.
RESULTS: A total of 23 studies were included. The mean age of presentation of C-ROCM was 54.6 years. The most common presentation was ptosis (72.7%), lid edema (60.6%), proptosis (60.6%), ophthalmoplegia (57.3%), loss of vision (53.7%), facial edema (34.7%), and nasal-blockage (11.8%). Evidence of intracranial spread was seen in 42.8% of cases. Rhizopus was the most common fungus (57.1%) isolated in fungal culture. Among C-ROCM patients, diabetes was the commonest comorbid condition, and the use of corticosteroids related to COVID-19 treatment was the most common risk factor (85.75%). Compared to controlled diabetics, C-ROCM was significantly higher among uncontrolled diabetics (odds ratio [OR] 0.15, 95% confidence interval [C.I.] 0.041-0.544, P = 0.0010). However, no significant association was seen between C-ROCM and COVID-19 severity (OR 0.930, 95% C.I. 0.212-4.087, P = 0.923). For treatment, amphotericin-B was the most common antifungal drug used which was followed by surgical options. However, mortality was high (prevalence 0.344, 95% C.I. 0.205-0.403) despite treatment.
CONCLUSION: Although local rhino-orbito symptoms were the first to appear, rapid intracranial extension was seen in a significant number of C-ROCM cases. Uncontrolled diabetes and excessive use of corticosteroid were the most common risk factors present among the C-ROCM cases. High index clinical suspicion is imperative (specifically among COVID-19 patients with diabetes), and routine screening may be helpful.

Introduction

Fungal infections are being recognized as important secondary infections among patients with COVID-19.[12] Pre-existing comorbidities, injudicious use of corticosteroids and antimicrobials, and lapses in infection control practices contributed to opportunistic fungal infection in COVID-19 cases.[3] Common systemic fungal infections reported among COVID-19 population are candidemia, invasive aspergillosis, and mucormycosis.[45] COVID-19–associated mucormycosis (CAM) is gaining significant attention across the globe.[67] The general population incidence of mucormycosis is low (0.005–1.7/million populations); however, a surge of increases in the number of cases has been reported in recent COVID-19 pandemic.[8]

In immunocompetent patients, the incidence of opportunistic infections is low in the maxillofacial region, which may be attributed to its high vascularity. However, mucormycosis is especially important in this region as they colonize in the nasal mucosa and invade the local defense mechanisms in this area and can lead to invasive infections.[9] Rhino-orbito-cerebral mucormycosis (ROCM) represents the most common form of mucormycosis involving the maxillofacial and orbital region, which may further spread intracranially in its advanced stage and is reported among COVID-19 patients.[10] Critically ill patients on oxygen therapy, mechanical ventilation, and prolonged intensive care unit stay are known risk factors of mucormycosis, especially C-ROCM patients,[10] and high rate of mortality was seen among those with intra-cranial extension.

Although ROCM is very much relevant in the context of COVID-19, and few systematic reviews have addressed this issue, till now, no meta-analysis is available to address the demographic and clinical profile, risk factors, impact of different interventions, and outcomes among C-ROCM cases. In this context, the current systematic review and meta-analysis was undertaken to address these knowledge gaps.

Materials and Methods

This meta-analysis was undertaken and reported in accordance with “Preferred Reporting Items for Systematic Review and Meta-Analysis guidelines”[11] (PRISMA) (PROSPERO registration I.D. CRD42021262943).

Objectives

Primary objectives

Clinical and demographic profile of C-ROCM patients

Microbiological profile of C-ROCM cases

Association between presence of risk factor/comorbidity and occurrence of C-ROCM

Evaluation of different intervention used in ROCM and their outcome (treatment success, treatment failure, and death).

Secondary objectives

Association between C-ROCM and diabetes (prevalence of diabetes among C-ROCM; both controlled and uncontrolled diabetes mellitus [DM])

Association of mortality and DM in C-ROCM

Association between C-ROCM and COVID-19 severity.

Inclusion criteria

Studies including diagnosed/confirmed mucormycosis patients, as defined by “Revision and Update of the Consensus Definitions of Invasive Fungal Disease From the European Organization for Research and Treatment of Cancer and the Mycoses Study Group Education and Research Consortium”[12] of all age groups associated with concurrent or post-COVID-19 infection

All types of observational study (prospective, retrospective, cohort, etc.) and case series reporting C-ROCM were included in our meta-analysis.

Exclusion criteria

Case reports, guidelines, consensus, editorials, and review articles were excluded from our meta-analysis.

Study population

Confirmed ROCM patients[12] including all age groups concurrent or post-COVID-19 C-ROCM.

Definitions

COVID-19 severity: COVID-19 severity definitions were adopted according to the World Health Organization “Therapeutics and COVID-19: living Guideline”[13] and patients were categorized into critical, severe, and nonsevere COVID-19 categories.[13]

Confirmed ROCM: Confirmed C-ROCM cases was defined according to the guidelines set by “Revision and Update of the Consensus Definitions of Invasive Fungal Disease From the European Organization for Research and Treatment of Cancer and the Mycoses Study Group Education and Research Consortium”[12]

Treatment success: Defined as response to treatment, i.e., stabilization of disease and disease-free patient[14]

Treatment failure: Death/mortality or progression of disease (showing radiological or clinical progression).[14]

Literature search and screening of articles

Eight literature databases were searched, i.e., PubMed, EMBASE, Google Scholar, Web of Science, Scopus, MedRxiv, BioRxiv, and SSRN from November 1, 2019, to June 30, 2021. Keywords used for data search were “Mucormycosis,” rhino, orbital, cerebral, ROCM, CAM, COVID-19, SARS CoV-2 without language restriction [detailed search strategy is given in Supplementary Table 1]. With completion of detailed search, duplicates were removed following which two authors (AB and PS) screened the articles/studies using the title and the abstracts of the study as predefined by the inclusion and exclusion criteria, following which full text of the relevant article was further evaluated for inclusion of our meta-analysis by using predefined inclusion and exclusion criteria. Any discrepancy among authors was resolved by consulting with BM.

Supplementary Table 1

Detailed search strategy

Key word Search:

1# ((((((((“mucormycosis”[All Fields]) OR (“mucor”) OROR (“mycoses”) [All Fields]) (“zygomycosis” [All Fields])) OR(“phycomycosis” [All Fields]) OR (“rhizopus”) [All Fields]) [AllFields]) OR (“absidia”) [All Fields]) OR (“cunninghamella”) [AllFields]) OR (“fungal infection”) [All Fields]) OR(“apophysomyces”) [All Fields] OR (“lichtheimia”) [All Fields] OR(“saksenaea”) [All Fields] OR (“rhizomucor”)2#(“rhino”[All Fields]) OR (“nose”) OR (“ENT”) OR (“Nasal”)3#(“orbital”[All Fields]) OR (“orbito”[All Fields]) OR (“orbit”[All Fields])OR (“Ocular”) OR (“Eye”) OR (“Ophthal*”)4#(“cerebral”[All Fields]) OR (“Brain”) OR (“Cranial”) OR (“Cranium”)5# 2# OR 3# OR 4#6# (“COVID-19”[All Fields]) OR (“sars cov 2”[All Fields])7# #1 AND #5 AND #6

ENT=Ear nose and throat

Data extraction

PS and AB independently extracted the data directly from full-length articles into structured tables. Collected data consisted of study country of origin, study type, risk factor or comorbidity, laboratory parameters, anatomic location of infection, diagnostic method, clinical manifestation, details of therapy used (antibiotic, steroid, antifungal use, surgery, etc.), and outcome of treatment (clinical and radiological response, laboratory parameters, hospital stay, treatment success and mortality etc.). In case of any study where our studied outcome parameters were not reported, corresponding authors were contacted for data via e-mail.

Quality assessment

The “methodological quality assessment” of the observational studies was assessed using “Newcastle–Ottawa Scale,”[15] and “risk of bias (ROB)” was assessed under three domains, i.e., selection, comparability, and outcome.[16] For assessing the quality of case series, scale mentioned by Murad et al.[17] was used (ROB domain evaluated: selection, ascertainment, causality, and reporting). Two investigators PS and AB independently evaluated the ROB of all the included studies. BM was consulted to solve the issue in case any discrepancy. Only high-quality studies were included in the meta-analysis.

Publication bias

”Publication bias” was evaluated by constructing a funnel plot and “Egger's regression test.”[18]

Statistical analysis

For single-group dichotomous data, pooled proportion/prevalence with 95% confidence interval (C.I.) was calculated. In case of dichotomous data having two groups, log odds ratio (OR) with 95% C.I. was used as a measure of association. In case of continuous data, for single-group analysis, pooled mean with 95% C.I. was calculated, and for two groups, mean difference with 95% C.I. was calculated. I2 and Chi2 statistics were used to assess heterogeneity.[19] In case of high heterogeneity (>50%), “random-effect model” was used for the analysis, else a “fixed-effect model” was used.[20] For the meta-analysis, “R software with Metafor Package”, R Foundation for Statistical Computing, Vienna, Austria” was used. Data conversion to different formats was done as per Wan et al., 2014.[21] Proportion was obtained as a ratio of total sample size. For prevalence, it was multiplied by 100.

Exploration of heterogeneity

For significantly high heterogeneity (>50%), etiology of high heterogeneity was investigated using “subgroup and meta-regression analysis.” Meta regression was conducted in case where the confounding variable was reported in 10 or more studies.[22]

Subgroup analysis

”Subgroup analysis” was done on the basis of study design (i.e., prospective vs. retrospective) and on the basis of studies including only severe and critical category COVID-19 population versus studies including COVID-19 population of all severity category.

Meta regression

Many factors (e.g., age, sex, comorbid condition, and therapy) may influence the mortality associated with C-ROCM.[23] In this context, “a meta-regression analysis” was done to observe the effect of these confounders on the final result, especially in the outcome of the disease (mortality).

Results

Details of included studies

After searching eight literature databases, we identified 827 articles. After removing duplicates, 615 studies were obtained, which were carried forward for further screening using title and abstract. Full-text screening was done for 67 relevant articles, out of which 23 studies fulfilling “predefined inclusion/exclusion criteria” were finally included in the systematic review and meta-analysis. The majority of included studies were from India (14 studies[672425262728293031323334] ) followed by 4 studies from Egypt,[35363738] 1 from Iran,[39] 1 from Turkey,[14] 1 from Chile,[40] 1 from French[41] and 1 from the Netherlands.[42] Among all these included articles, 4 were preprint[31333441] [detailed characteristics of included studies showed in Supplementary Table 2 and Supplementary Figure 1 shows PRISMA flow diagram].

Supplementary Table 2

Details of included studies

Author, year/country/study design time/duration of study	Study population/sample size (n)	Age group (years)	COVID-19 category/T	Presenting clinical feature/COVID-19 marker at presentation
Deepti Satish[1] 2021/IndiaRetrospective studyMarch-December 2020 were included	Patients with invasive fungal infectionn=25COVID-19 (+) =11	30-74	Asymptomatic: 2Mild: 2Moderate: 3Severe: 4Concurrent=8Post-COVID-19=3	Unilateral facial swelling, retro-orbital pain, ptosis, and headache
Mishra et al.[2] 2021/IndiaRetrospective studyAugust-December 2020	COVID-19-associated ROCM patientn=10	37-78	Mild=3Moderate=6Severe=1	Ocular pain=1Loss of vision=2Chemosis=2Epistaxis=1Facial pain=1Headache=1Ocular swelling=1Nasal block=2Nasal crusting=1
Sen et al.[3] 2021/IndiaRetrospective studyAugust-December 2020	COVID-19-associated ROCM casesn=6	60.5 (range 46.2-73.9)	Moderate-to-severe COVID-19=5Post-COVID-19 cases=5Concurrent cases=1T=30-42 days	Initial presentation: pain, redness, periocular swelling=6Developed ptosis, ophthalmoplegia, painful loss of vision average 2 days from day of presentation
Sarkar et al.[4] 2021/IndiaCase seriesOctober-November 2020	A cluster of clinically diagnosed orbital mucormycosis with concurrent COVID-19 illness over 2 months n=10	45.5	Mild-to-moderate: 1Severe: 9	CRAO=6Loss of vision=4DKA=4
Moorthy et al.[5] 2021/IndiaRetrospective observational multicenter studyMay 2020-December 2020	Patients with aggressive maxillofacial and ROCM cases concurrent or post COVID-19 positiven=18	54.6 (35-73)	No detailConcurrent case=4Post COVID-19=14	Loss of vision=12/18maxillary sinusitis=18/18Maxillary necrosis=14/18Intracranial extension=9/18
Sharma et al.[6] 2021/IndiaProspective observational studyAugust-December 2020	Mucormycosis of PNS with history of COVID-19n=23	NA	NAConcurrent COVID-19 infection: 4Post COVID-19 mucormycosis: 14	Intraorbital extension=43.47%Intracranial extension=8.69%Involvement of ethmoid sinus=100%Classical black escher on hard palate=9/23
Nehara et al.[7] 2021/IndiaCase seriesNovember-December 2020	COVID-19-associated ROM patientn=5	62.2	Severe to critical=2Mild-to-moderate=3Concurrent COVID-19=3Post COVID-19=2	Loss of vision=3Headache=3Ptosis=2Proptosis=4Palate inv=3Complete ophthalmoplegia=2Epistaxis=3
Ravani et al.[8] 2021/IndiaRetrospective studySeptember 2020-mid-March 2021	All biopsy-proven mucormycosis patientTotal n=31COVID-19=19All are post-COVID, no case of concurrent COVID-19	56.3	NAT=average 2 monthsNo concurrent COVID-19 case	Commonest presentation: Diminution of vision (<6/60 in 80.64%), ophthalmoplegia (77.4%)
Diwakar et al., 2021[9]/IndiaCase seriesPreprint	CAM patient of pediatric age group with type 1 DMn=2	Case 1=11Case 2=13	Asymptomatic COVID-19 infectionROCM developed during course of DKAT=3 days (case 1)T=20 (case 2)	Case 1: Weight loss, unilateral pain and swelling of eye, high-grade fever, complete ophthalmoplegiaCase 2: Unilateral pain, swelling and diplopia, complete ophthalmoplegia, nonreactive pupil
Gangneux et al., 2021[10]/FrenchNational multicenter observational study 18 French ICUPreprintFebruary 29 and July 9, 2020	Severe COVID-19 population with IFI≥18 years age groupn=57	59.4±12.5	Severe COVID-19=57	Mucormycosis prevalence 1.2%
Buil et al., 2021[11]/NetherlandsCase series	Secondary fungal infections caused by Mucorales species in COVID-19 patientsn=4	>50	3 CAM cases occurred in ICU one outside ICU	Pulmonary, rhino-orbital cerebral and disseminated infection
Fouad et al., 2021[12]/EgyptRetrospective observational studyMarch 25, 2020-September 25	ROCM patientsn=12COVID-19 positive=6Concurrent case=5Post-COVID-19 positive=1	51.2 (16.7)	Moderate-to-severe=5	Orbital invasion=allCerebral invasion=8Lid edema(50%), conjunctival chemosis (50%) diminution of vision (41.7%), proptosis (33.3%), facial edema (25%), nasalCrusts (25%), total ophthalmoplegia (16.7%)Paralytic esotropia (8.3%)
El-Kholy et al., 2021[13]/Egypt/longitudinal prospective studyAugust 2020-December 2020	COVID-19-associated acute invasive fungal rhino sinusitisn=36	52.92±11.30	Mild=11Moderate=13Severe=12MV needed=3	Headache and facial pain (75%)Facial numbness (66.7%)Ophthalmoplegia, and visualLoss (63.9%)Proptosis=19Diplopia=6Palatal involvement=11.1%Necrosis of cheek=1Orbital apex syndrome=3
Wael F. Ismaiel et al., 2021[14]/Egypt/retrospective studyJanuary 2017-December 2020	Diagnosed AIFRSn=56Total post COVID-19 cases=18	58.38±12.2	ARDS=3	-
Manar M. Ashour, et al., 2021[15]Case seriesMay 2020-February 2021	Confirmed COVID-19AIFSn=7	-	T=12-35 days	Right upper eyelid edema=4Total ophthalmoplegia=5No PL=5, black nasal crusts=6Rapidly deteriorating visual acuity=1Conjunctival chemosis=1Proptosis=1Paralytic esotropia=1Bilateral oroantral fistulae=1Eyelid edema=1Reduced visual acquity=1Facial edema=2Skin discoloration=1bilateral Panophthalmitis=1, necrosed nasal septum=1Nacrosed hard palate=1Painful orbital proptosis=1Drop of vision=1Panophthalmitis=1Septal ulceration=1Panophthalmitis=4Orbital compartment syndrome=1Optic nerve ischemia/inflammation=2Bilateral orbital panophthalmitis=2Intracranial complications were: Perineural spread=6Cavernous sinus involvement=6Meningeal/epidural infiltration=3ICA vasculitis/thrombosis=4 ICA mycotic aneurysm=intracerebralAbscess=2 and cerebral infarctions=3
Ricardo Rabagliati et al.[16]/2021/Santiago, ChileRetrospective case seriesDuring May 18-July 18, 2020	Nonimmunocompromised adult CAIMI cases with severe COVID-19 cases admitted in ICUTotal severe to critical COVID-19 cases=146 n=16	Median age65 (30-89)	All severe to critical COVID-19 casesT=18.5 (range 1-47) days 14.5 days (range 0-28)After ICU admission 12.5 days (range 0-28) after IMV	-
Bayram et al., 2021[17]/prospective observational studyMarch 2020-December 2020	COVID-19-associated ROM cases in severe COVID-19 patientsn=11	-	All severe COVID-19 casesT=5.1±1.8 days (2-8 days)Increased D-dimer=all patientLow molecular weight heparin received=all mean D-dimer level=1362.4±468.9 μg/L	Proptosis (100%)Ophthalmoplegia (63.6%), orbital pain (81.8%), conjunctivalHyperemia/chemosis (81.8%), ptosis (63.6%), fixed and dilated pupil (63.6%), vision loss (63.6%), endophthalmitis (54.5%), decreased vision (27.3%), orbital apex syndrome 7 (63.3%) syndrome, orbital cellulitis (36.4%)
Joshi et al., 2021[18]/India/retrospective study March 1, 2021-April 15, 2021	COVID-19 and invasive ROCM cases	55.2±13, range 34-76	Severe to critical COVID-19	-
Atul Patel et al., 2021[19]/India/multicenter retrospective study in 16 healthcare centers across IndiaSeptember-December 2020	Confirmed mucormycosis cases with and without COVID-19Total mucormycosis patient=287CAM=187 (65.2%)	53.4 (SD 17.1)	T=median time 18 days (IQR 11-27)	Commonest is rhino-orbital mucormycosis (58.2%), followed by ROCM, pulmonary and other sites
Muley et al., 2021[20]/India/observational studyApril 15, 2021, and May 15, 2021 Preprint	Probable RMMn=30	-	23 (76.6%) gave a history of hospitalization due to COVID-1917 (23.3%) had no history of hospitalization	6.6% of cases were in Grade 1, while 53.5% were in Grade 2, and 40% were in Grade 3
Farzad Pakdel et al., 2021[21]/Iran/cross-sectional descriptive multicenter study in five COVID-19 hospitalized centers in Tehran, IranApril-September 2020	Biopsy-proven ROCM patients with COVID-19n=15	52 (range 14-71)	Mild=3Moderate=5Severe=7T: 1-37 days	Unilateral periorbital pain and edema (73%), ptosis (73%), acute vision loss (73%), proptosis (67%), unilateral facial edema (60%), cranial nerve palsy (60%), headache (33%), fever (27%), nasal blockage (13%) and ear pain (7%)
Rajalakshmi Arjun et al., 2021[22]/India/case series	COVID-19-associated rhino-orbital mucormycosisn=10	53.0±12.1	All cases had past H/O COVID-19T=17.0±3.6 days	Headache=6Facial swelling=3Facial numbness=1Lid swelling=4Black discoloration of palate=1Ocular pain=3Earache=1Diplopia=1Sphenoid=8Intracranial spread=3Optic nerve involvement=2
Sebastian et al., 2021[23]/India/case series	CAIFSn=3	61±2.64	Severe to critical=3Bilateral COVID-19 pneumonia with ARDS=3Altered sensorium=1RF and shock=1	Nasal blockage=1facial swelling=1Periorbital swelling=2Blackening of middle turbinate=2Blackening of the lateral wall of nose with crusts and discharge=1Thick dirty nasal discharge=1Eye pain=1Restricted eye movements=1Diminution of vision=1Cavernous sinus thrombosis=1Proptosis=1Periorbital bluish discoloration=1
Author, year/country/study design time/duration of study	Co-morbidities/risk factor	Hyperglycemia at presentation	Treatment	Comment
Deepti Satish[1] 2021/IndiaRetrospective studyMarch-December 2020 were included	Immunocompromised-23Immunocompetent=2Majority of the patients with DM and HTN with underlying IHD and CKDLeukemia=1	CAM cases had HbA1c level 7-15 mg/dl (majority >10 mg/dl)COVID-19 negative Mucor patients had HbA1c level 6-13 mg/dl (majority <10 mg/dl)	Empirical IV Amph B in severe CAM patients and debridement done once stableCOVID-negative cases: Debridement followed by IV AmphBIV AmphB dose: 50 mg/kg/day with cumulative dose of 1.5-2 gPostoperatively local diluted AmphB douching all casesDebridement: 20	Details of COVID-19 patient profile is not availableRedebridement done in 1 patient after 1 month due to vision lossSignificant delay in the surgical management of moderate-to-severe COVID-19 patients were there due to lack of fitness for GA
Mishra et al.[2] 2021/IndiaRetrospective studyAugust-December 2020	DM=8HTN=3CKD=2CLD=1IHD=1Hypothyroidism=2	No detail available	Steroid=6Amph B=10Posaconazole=1RDV=6Tocilizumab=1	4/10 patients were diagnosed and managed for COVID-196/10 were treated for COVID-19 outsideLost to follow-up=1
Sen et al.[3] 2021/IndiaRetrospective studyAugust-December 2020	DM (Type 2): 100%2/6 were diagnosed of DM with the onset of COVID-19	Average FBS=222.5±144.4 (86-404) mg/dL at presentationMean duration of DM=5.9±4.9 (median 6.5) yearsUncontrolled DM=5DKA=3Case 2: Moderate NPDR with DME	Antifungals started after microbiological confirmationLipAmpB B (5 mg/kg/day, up to a maximum of 10 mg/kg/day I.V. for CNS infections)Oral posaconazole (loading dose 300 mg BD day 1 followed by maintenance dose 300 mg OD)Orbital-exenteration: In suboptimal response to systemic antifungals within 72 h	Male cases 100%All were referred cases
Sarkar et al.[4] 2021/IndiaCase seriesOctober-November 2020	DM: 10/10 (100%)Steroid: 10/10 (100%)	DKA at admission: 4/10DKA during COVID-19 therapy: 5/10	All patient received dexamethasone as per NIH guidelineLipAmpB: All casesRDV=4/10Ventilatory support=9/10	Detail dose not given
Moorthy et al.[5] 2021/IndiaRetrospective observational multicenter studyMay 2020-December 2020	DM: 16/18	All poorly controlled DM	Mucormycosis: LipAmpB at 3-5 mg/kg to a cumulative dose of 3-5 gAspergillosis: Oral voriconazole (6 mg/kg IV 12 hourlyDay 1/4 mg/kg IV 12 hourly subsequently) and posaconazole (200 mg)Orbital exenteration: 7Maxillectomy: 11Endoscopic sinus surgery: 17	Maxillary necrosis=14/18, which is statistically significant (P=0.03)Lost to follow up=1
Sharma et al.[6] 2021/IndiaProspective observational studyAugust-December 2020	DM: 21HTN: 14Renal failure: 1Steroid user=100%	DM: 21Uncontrolled DM: 12 (HbA1c>6.5%)	Intra-orbital AmpB: 2	None gave consent for orbital exenterationMortality: 0%
Nehara et al.[7] 2021/IndiaCase seriesNovember-December 2020	DM: 5HTN: 2	Uncontrolled DM=3DKA=1Ketosis without acidosis=1	LipAmpB=allPosaconazole=1MV=1Broad-spectrum antibiotics=allIonotrops=2	-
Ravani et al.[8] 2021/IndiaRetrospective studySeptember 2020-mid-March 2021	COVID-19 positivity: 61.2%Uncontrolled DM: 96.7%Steroid user: 61.2%HTN=54.8%IHD=3.22%Kidney disease=6.45%	Uncontrolled type 2 DM=29 (93.54%)Uncontrolled type I DM=1Newly diagnosed DM=6Average duration of DM=4.4 yearsAverage HBA1C=7.57 mmol/molDuration of steroid=7-14 days	IV LipAmpB B: 31 (100%) after microbiological confirmationSinus debridement: 31 (100%)Exenteration: 4 (12.9%)IV LipAmpB dose=3-5 mg/kg body weight/dayOral posaconazole: Add-on therapy at time of resolution (5 mg/kg body weight/day)	Cerebral involvement and HbA1c ≥8 found to be significant in prediction of 75-day mortality (P≤0.05)
Diwakar et al., 2021[9]/IndiaCase seriesPreprint	Type 1 DMNO H/OSteroid exposure	Both are cases pf type 1 DM	I.V. LipAmpB=2Retro orbital AmpB=2Craniotomy (for brain abscess)=2Orbital exenteration with resection of the involved sinuses. =2I.V LipAmpB dose (10 mg/kg/day) in 5% dextrose over 1 h for 6 weeks	Case 1: hypokalemia (serum K+=2.8 mmol/L), hyperglycemia 96 (RBS=329 mg/dLCase 2: LFT: Hypoalbuminaemia 130 (3 g/dL) and low albumin/globulin ratio (1.00)
Gangneux et al., 2021[10]/FrenchNational multicenter observational study 18 French ICUPreprintFebruary 29 and July 9, 2020	DM: 32·9%HTN: 50·1%Lymphopenia at presentation: 64·2%Mean duration of MV: 27·1±19·8 days	-	Detail for mucormycosis patients not available	Detail profile of for mucormycosis patients not available
Buil et al., 2021[11]/NetherlandsCase series	DM=2Steroid=4MV=4CLL=1Obesity=1	DM=2Uncontrolled DM=1	Tocilizumab=1LAmpB=4Prosaconazole=1Voriconazole=1Isavuconazole=2INF Gama=1	Microscopy showed nonseptate hyphae in the patient’s urine in 1 case died 3 out of 4 cases
Fouad et al., 2021[12]/EgyptRetrospective observational studyMarch 25, 2020-September 25	DM=10Hematological malignancy=2 received chemotherapyCKD=3IHD=1	DM=10Uncontrolled DM=10/12 cases (83.3%)Occurrence of DM at time of ROCM diagnosis=9Median HbA1c=9.7%DKA=1	Debridement done=7 (58.3%)	-
El-Kholy et al., 2021[13]/Egypt/longitudinal prospective studyAugust 2020-December 2020	DM (27.8%)HTN (16.67%)Associated malignancy, uncontrolled leukemia, and pancreatic cancer=5.56%	DM=10 (27.8%)3/10 COVID-19-associated due to corticosteroid therapy	Antifungal therapy and surgery=34Only antifungal therapy=2LipAmpB=26Voriconazole=8Both=2Posaconazole=3	Individual data for COVID-19-associated mucormycosis not givenEarly cases (n=7) with limited sinonasal involvement showed better results
Wael F. Ismaiel et al., 2021[14]/Egypt/retrospective studyJanuary 2017-December 2020	DM=44.4% of post-COVID-19 AIFRSImmunosuppressive=16.7%HTN=10 obesity=8Smoking=12Allergic rhinitis=14Asthma=4 COBD andCardiac diseases=5Otitis media=6Renal dysfunction=3Liver dysfunction=2Thrombocytopenia and leucopenia=4Immunosuppressive drugs=3Antibiotic or antiviral therapy=12Steroid use=7Chemotherapy=1	DM=44.4% of post-COVID-19 AIFRS	-	Incidence of AIFRS is more prominent in post-COVID-19 patients than in non-COVID-19 especially in immunocompromised patients, diabetic, renal and liver dysfunction patients and patients with risk factors for rhino sinusitis
Manar M. Ashour, et al., 2021[15]Case seriesMay 2020-February 2021	DM (n=6), HTN (n=2), end stage renal disease (n=2), hyperlipidemia (n=2), ischemic heart disease (n=1), previous cerebral stroke (n=1)	-	AmpB=7Itraconazole=1Ambisome=1Surgical debridement=7	Patients had radiologic features of aggressive late-stage forms of the disease process with a consequent long-term morbidity rate of 100% and a high mortality rate of 37.5%
Ricardo Rabagliati et al.[16]/2021/Santiago, ChileRetrospective case seriesDuring May 18-July 18, 2020	HTN=9 (56.3%)Asthma/COPD=4 (25%)DM=4 (25%)Obesity=3 (18.8%)Median worst PaO2/FiO2 for each patient was 124 (range 57-476)None were immunocompromised	-	Antifungal therapy=13 (81.3%)Voriconazole=10 (76.9%)LipAmpB=2 (15.4%)Isavuconazole=1 (7.7%)Therapeutic drug monitoring received with patients under voriconazole therapy=9 patients receiving (median 3.9 mg/L; range 0.1-7.2 mg/L)	Severe hypoxia, broad-spectrumAntibiotics, high corticosteroid doses, prolongedICU stay, long intubation period, and airway/lung damage and infarction areas are predisposing condition for critically ill patients
Bayram et al., 2021[17]/prospective observational studyMarch 2020-December 2020	COVID-19-associated ARDS=11Corticosteroid=11DM=8 (72.7%)CRF=3ARF=2Hematological malignancy=1 under immunosuppressive	Type 2 DM=8 (72.7%), all uncontrolledMean duration of diagnosis=12.1±4.4 years (range: 7-19 years)	IV and retro bulbar LipAmpB=all casesIntravitreal LipAmpB in patients with endophthalmitisIV dose=1.0 mg/kg/day, increasing a total dose of 2.5-3 gRetro bulbar dose=1 ml of 3.5 mg/mlIntravitreal dose=5 μg/0.1 mlSecond surgical procedure=8 (72.7%)Mean number of retrobulbar and intravitreal AmpB injections=2.2±0.6 and 2.3±0.5Interval between repeat injections varied from 2-8 days	Diagnosis of mucormycosis was made in all patients during the COVID-19 treatmentThe mean MIC of amphotericin B was 2.5±1.0 μg/ml (range: 1.5-4 μg/ml)The MIC of voriconazole was >32 μg/ml in all patientsAll endophthalmitis cases showed posterior scleritis, two developed retinochoroiditis followed by retinoschisis, and one had corneal edema
Joshi et al., 2021[18]/India/retrospective studyMarch 1, 2021-April 15, 2021	DM=22Uncontrolled DM=13HIV=2MV received=2Corticosteroids=2 (duration, 10-14 days)	-	Amphotericin B=allSurgical debridement=10Orbital exenteration=10 s	-
Atul Patel et al., 2021[19]/India/multicenter retrospective study in 16 healthcare centers across IndiaSeptember-December 2020	COVID-19 only 61 (32.6%)Glucocorticoids 48/61 (78.7)DM=113 (60.4%)Posttrauma=3 (1.6%)Hematological malignancy=2 (1.1%)Renal transplantation=3 (1.6%)Median cumulativeDexamethasone-equivalent dose=84 mg (range 18-1,343 mg)Tocilizumab=5 (2.7%)Hypoxemia due to COVID-19 and inappropriate glucocorticoid administration were associated with development of late CAM	Uncontrolled DM=62.7%DKA=early (28%), late CAM (5%)Newly diagnosed DM=39/187 (20.9%)COVID-19 was the only underlying disease in 61/187 (32.6%)	LipAmp B=136 (72.7%)AmpD deoxycholate=31 (16.6%)Posaconazole=73 (39.0%)Isavuconazole=19 (10.2%)Single antifungal drug=95 (50.8%)Concurrent=13 (7.0%)Sequential=79 (42.5%)Combined medical and surgical therapy=131 (70.1%)LipAmpB dose=5 mg/kg 1×/d for 4-6 weeksAmpB deoxycholate=1 mg/kg 1×/d for 6-8 weekSequential antifungal drug treatment improved mucormycosis survival	CAM prevalence=0.27% among hospitalized COVID-19 patientsNewly diagnosed DM in CAM patients were significantly higher (P=0.02)DKA seen more in non-CAM patients (16/187 [8.6%]) than in CAM patients (27/100 [27%]; P=0.0001)COVID-19-related hypoxemia and improper glucocorticoid use independently were associated with CAM
Muley et al., 2021[20]/India/observational studyApril 15, 2021, and May 15, 2021 Preprint	76.66% had a history of COVID-19, and 66.66% had hospitalization and steroid administration history	84% patients had diabetes mellitus	36% were operated on by OMFS, and 6.6% were referred to an ENT specialist	-
Farzad Pakdel et al., 2021[21]/Iran/cross-sectional descriptive multicenter study in five COVID-19 hospitalized centers in Tehran, IranApril-September 2020	DM=13 (86%)HTN=7 (46%)Hematologic malignancies=2 (13%), Asthma 2 (13%)Cardiovascular disease 2 (13%), hepatic cirrhosis 1 (6%), hypothyroidism 1 (6%), TB 1 (6%), immunosuppressive therapy 7 (46%), chemotherapy 2 (13%), neutropenia 3 (20%), ketoacidosis 1 (6%)	46.6% patients previously received intravenous corticosteroid therapy	IV AmpB: All casesOral posaconazole: 4 (27%)Combined anti-fungal therapy: 6 (40%)IV AmpB (5 mg/kg daily for 4-6 weeks)Orbital exenteration=5Palatal debridement=2	Anti-fungal combination therapy was significantly associated withBetter outcome (P=0.003)
Rajalakshmi Arjun et al., 2021[22]/India/case series	DM=allHTN=2Hypothyroidism=1CAD=3CKD=1RA=1Supplemental O2 required=8Steroid therapy=8Broad-spectrum IV antibiotics=9Steam inhalation: 9	DM=all casesMean HbA1c: 10.2±2.0	Endoscopic sinus surgery and debridement was done in all patients within 24 hAntifungals used included liposomal amphotericin BAmphotericin B deoxycholate and isavuconazole	-
Sebastian et al., 2021[23]/India/case series	DM=1CABG=1CKD=1HTN=1MV=2 (Case 1=on mechanical ventilation for 7 days)Peptic ulcer=1BSA=3Steroid=3Renal replacement therapy=2	All three are diabetic controlled	Case 1=liposomal amphotericin B was given (total dose of 3050 mg). Subsequently he was continued on voriconazoleSurgical debridement was doneCase 2=total dose of 850 mg of liposomal amphotericin B	Case 1: Although clinical resolution was seen, during the post-COVID recovery phase patient developed myocarditis with cardiac arrhythmia and expiredCase 2=surgical debridement could not be carriedCase 3=developed massive peptic ulcer bleed and went into irreversible shock and died

MV=Mechanical ventilation, IMV=Invasive mechanical ventilation, ARDS=Acute respiratory distress syndrome, ROCM=Rhino-orbito-cerebral mucormycosis, CAM=COVID-19 associated mucormycosis, T=Time to develop mucormycosis from COVID-19 diagnosis, IHD=Ischemic heart disease, CKD=Chronic kidney disease, RDV=Remdesivir, FBS=Fasting blood sugar level, DKA=Diabetic ketoacidosis, NPDR=Nonproliferative diabetes mellitus, DME=Diabetic macular edema, CRAO=Central retinal arterial occlusion, PNS=Paranasal sinus, BSA=Broad spectrum IV antibiotics, AIFRS=Acute invasive fungal rhino sinusitis, RMM=Rhinomaxillary mucormycosis, ICU=Intensive care unit, OMFS=Oral and maxillofacial surgeons, PL=Perception of light, IFI=Invasive fungal infections, SD=Standard deviation, IQR=Interquartile range, NA=Not available, DM=Diabetes mellitus, HTN=Hypertension, ARF=Acute renal failure, COPD=Chronic obstructive pulmonary disease, CRF=Chronic respiratory failure, CABG=Coronary artery bypass grafting, HbA1c=Hemoglobin A1c, RBC=Red blood cell, MIC= Minimal inhibitory concentration, LFT= Liver function test, NIH=National institute of health, ENT= Ear nose and throat, ICA= Internal carotid artery RF=Respiratory failure, TB= Tuberculosis, RA= Rheumatoid arthritis, GA= General anaesthesia, ROM=Rhino orbital mucormycosis, AIFS= Acute invasive fungal sinusitis, CAIMI=Covid-19 associated invasive mucorales infection

The Egger's test shows no evidence of publication bias with P = 0.633. [Supplementary Figure 2].

Primary objective

Demographic profile of patients with COVID-19–associated rhino-orbito-cerebral mucormycosis

Mean age at presentation

A total of 17 studies[67142526272931323435373940414243] reported mean age of presentation of C-ROCM. The mean age of presentation of C-ROCM was 54.6 years (95% C.I. 44–65, I2 = 99%, random-effect model) [Figure 1]. In the pooled data, it was seen that Bayram et al.[14] and Diwakar et al.[31] reported extremes of age group (mean age 73 years and 12 years, respectively) as compared to other studies. Hence, a sensitivity analysis was conducted by excluding these two studies; however, the mean age of presentation remained similar (56.8 years [95% C.I. 54–58]) and heterogeneity reduced to 50% (I2 = 50%, fixed-effect model) [Supplementary Figure 3].

Figure 1

Mean age at presentation in C-ROCM patients

Sex ratio

Males were the more frequently affected gender (17 studies,[6714242526272829313334353739404243] prevalence 0.784, 95% C.I. 0.721–0.846, I2 = 28.96%, fixed-effect model).

Clinical presentation of COVID-19–associated rhino-orbito-cerebral mucormycosis population

Mean time period of presentation of C-ROCM from the time of COVID-19 diagnosis was 20 days (10 studies,[7141426293435404243] 95% C.I. 15–24 days) [Supplementary Figure 4]. There was no significant difference in odds of occurrence of ROCM among patients with concurrent COVID-19 infection versus post-COVID-19 infection (7 studies,[6262829303742] OR [log scale] 0.306, 95% C.I. 0.042–2.246, P = 0.244, I2 = 79.46%, random-effect model) [Supplementary Figure 5].

Regarding ocular presentation, unilateral involvement was commonest, while bilateral ocular involvement was reported only in small number of patients (2/12, i.e., 16.7%, single study[37] ). The most common ocular presentations were lid edema (9 studies,[252629313435373943] prevalence 0.606, 95% C.I. 0.385–0.826), proptosis (8 studies,[1426293537383943] prevalence 0.606, 95% C.I. 0.385–0.826), ophthalmoplegia (8 studies,[1429303135373843] prevalence 0.573, 95% C.I. 0.395–0.751), loss of vision (10 studies,[6142526272829353839] prevalence 0.537, 95% C.I. 0.282–0.793), facial edema (6 studies,[293435373943] prevalence 0.347, 95% C.I. 0.216–0.477), and ptosis (4 studies,[14262939] prevalence 0.727, 95% C.I. 0.534–0.920). Pupil involvement was reported by 3 studies[142631] (prevalence 0.581, 95% C.I. 0.362–0.801), while optic nerve involvement was reported by 2 studies[3435] (prevalence 0.230, 95% C.I. 0.031–0.430). Rare but vision-threatening presentations included panophthalmitis (single study,[35] n = 7, unilateral in 4/7 and bilateral in 1/7), endophthalmitis (single study,[14] 6/11, 54.5%), orbital compartment syndrome (single study,[35] 1 out of 7, 14.2%), and central retinal artery occlusion (single study,[27] 6 out of 10, 60%).

Regarding rhino/nasal presentation the most common presenting features reported were nasal blockage (2 studies,[2939] n = 25, overall 3 out of 25, prevalence 0.118, 95% C.I. −0.008–0.244), black nasal discharge (single study,[29] n = 5, 2 out of 5), and epistaxis (2 studies,[2529] n = 15, 3 out of 15).

Palatal involvement was also reported by few studies (3 studies,[262829] n = 21, overall 13 out of 21, prevalence 0.575, 95% C. I. 0.162–0.989). “Classic black Eschar” has been reported in small number of patients (single study,[28] n = 23, 9 out of 23), and bleeding gum (single study,[25] n = 10, 1 out of 10) has also been reported as infrequent presentation [detail clinical presentation and their prevalence in C-ROCM patients is summarized in Table 1].

Table 1

Pooled prevalence of proportion of various clinical presentations described among COVID-19-associated rhino-orbito-cerebral mucormycosis patient population

Clinical feature	Number of studies/reference	Event/total	Pooled data		SE	Over all P	Heterogeneity
			Pooled proportion	95% CI			I2 (%)	Tau 2/Q/DF	Heterogeneity P
Clinical feature
Loss of vision	10[6,14,252627,29,35,38,39,44]	71/141	0.537	0.282-0.793	0.130	<0.001	93	0.151/147.530/9	<0.001
Lid edema	9[25,26,29,31,34,35,37,39,43]	35/70	0.515	0.278-0.751	0.121	<0.001	84	0.105/52.026/8	<0.001
Ophthalmoplegia	8[14,293031,35,37,38,43]	66/107	0.573	0.395-0.751	0.091	<0.001	72	0.042/25.310/7	<0.001
Proptosis	8[14,26,29,35,373839,43]	56/95	0.606	0.385-0.826	0.112	<0.001	87	0.082/56.483/7	<0.001
Facial edema	6[29,34,35,37,39,43]	19/52	0.347	0.216-0.477	0.067	<0.001	9	0.003/5.514/5	<0.001
Palatal involvement	3[26,29,44]	13/21	0.575	0.162-0.989	0.211	0.006	79	0.104/9.717/2	0.008
Ptosis	4[14,26,29,39]	26/37	0.727	0.534-0.920	0.099	<0.001	53	0.020/6.458/3	0.091
Pupil involvement	3[14,26,31]	11/19	0.581	0.362-0.801	0.112	<0.001	0	0.000/0.356/2	0.837
Palatal Eschar	6[26,29,34,35,38,44]	20/74	0.356	0.058-0.654	0.152	0.019	91	0.120/59.771/5	<0.001
Optic nerve involvement	2[34,35]	4/17	0.230	0.031-0.430	0.102	0.023	0	0.000/0.163/1	0.687
Nasal blockage	2[29,39]	3/25	0.118	−0.008-0.244	0.064	0.067	0	0.000/0.067/1	0.796
Radiological feature
Sinonasal involvement	4[24,38,39,43]	47/79	0.549	−0.013-1.11	0.287	0.056	98	0.319/235.393/3	<0.001
Rhino-orbital involvement	8[7,14,26,31,37,39,43,44]	165/259	0.708	0.539-0.877	0.086	<0.001	87	0.044/57.998/7	<0.001
Intracranial spread	13[6,7,14,24252627,29,32,35,37,39,44]	103/333	0.428	0.297-0.560	0.067	<0.001	82	0.042/69.656	<0.001
Involvement of orbital apex	2[14,34]	8/21	0.257	−168-0.882	0.268	0.183	89	0.129/9.578/1	0.002
Involvement of cavernous sinus	8[26,29303132,373839]	36/132	0.403	0.22-0.595	0.098	<0.001	88	0.059/60.787/7	<0.001

CI=Confidence interval, SE=Standard error

Radiological involvement

Radiological investigation, especially computed tomography (CT) scan and magnetic resonane imaging (MRI), has been used as a diagnostic tool to evaluate the extent of involvement of mucormycosis infection almost all the authors. Our meta-analysis showed the most common involvement to be rhino-orbital involvement (8 studies,[714262831373943] prevalence 0.708, 95% C.I. 0.539–0.877), followed by sinonasal involvement (4 studies,[24383943] prevalence 0.549, 95% C.I. −0.013–1.11) and intracranial involvement (13 studies,[671424252627282932353739] prevalence 0.428, 95% C.I. 0.297–0.560). Involvement of orbital apex (2 studies,[1434] prevalence 0.257, 95% C.I. −168–0.882) and cavernous sinus was seen in small number of patients (8 studies,[2629303132373839] prevalence 0.403, 95% C.I. 0.22–0.595).

Regarding paranasal sinus involvement, “ethmoid sinus” was the most common paranasal sinus involved which was reported to be 80% (n = 10, 8 out of 10) by Arjun et al.,[34] 90.9% (n = 11, 10 out of 11) by Bayram et al.,[14] and 100% (n = 23, 23 out of 23) by Sharma et al.[28] Pansinusitis was seen in 60% (n = 5, 3 out of 5) by Nehara et al.,[29] 77.4% (n = 31, 24 out of 31) by Ravani et al.,[30] 90.9% (n = 11, 10 out of 11) by Bayram et al.,[14] and 100% (n = 6, 6 out of 6) patients by Sen et al.[26] Pakdel et al.[39] reported that the most common sinus involvement was pansinusitis in their study population.

Regarding orbital involvement, isolated orbital involvement was seen in 13% of cases (n = 15, 2 out of 15) by Pakdel et al.[39] Extraocular muscle involvement was seen in 19.35% of cases (n = 31, 6 out of 31) and orbital cellulitis in 61.29% (n = 31, 19 out of 31) by Ravani et al.[30] Subperiosteal abscess was seen in 5.6% of patients (n = 36, 2 out of 36) by El-Kholy et al.[38]

Regarding intracranial involvement, transverse and sigmoid sinus thrombosis was seen in 2.7% (n = 36, 1 out of 36), temporal lobe abscess in 2.7% (n = 36, 1 out of 36), encephalitis in 5.4% (n = 36, 2 out of 36), and central retinal artery thrombosis in 5.4% (n = 36, 2 out of 36) by El-Kholy et al.[38] Cavernous sinus thrombosis was seen in 60% (n = 5, 3 out of 5) patients by Nehara et al.[29] and 3.22% (n = 31, 3 out of 31) by Ravani et al.[30] Internal carotid artery thrombosis was seen in 6.45% (n = 31, 2 out of 31) and 16.7% (n = 25, 4 out of 25) by Ravani et al.[30] and Fouad et al.,[37] respectively. Brain infarct was seen in 20% (n = 5, 1 out of 5) patients by Nehara et al.[29] while cerebral abscess was seen in 16.7% (n = 25, 4 out of 25) population by Fouad et al.;[37] meningitis was seen in 8% (n = 25, 2 out of 25) by Joshi et al.;[32] 5th nerve involvement in 4% (n = 25, 1 out of 25) patients; and black turbinate sign in 32% cases (n = 25, 8 out of 25) in a single study.[32]

Very small number of studies reported pulmonary involvement and disseminated form of mucormycosis in C-ROCM patients. Buil et al.[42] reported bilateral consolidations in 25% patients (n = 4, 1 out of 4) and pulmonary cavities and a reversed halo-sign in 25% patients (n = 4, 1 out of 4). Rabagliati et al.[40] reported pulmonary cavitations in 6.25% (n = 16, 1 out of 16), pulmonary nodules in 6.25% (n = 16, 1 out of 16), both cavitations and nodule in 6.25% (n = 16, 1 out of 16), pleural effusion in 12.5% (n = 16, 2 out of 16), and pneumothorax and bullas in 12.5% patients (n = 16, 2 out of 16). Patel et al.[7] reported pulmonary involvement in 8.6% (n = 187, 16 out of 187) of C-ROCM patients. Sebastian et al.[43] reported bilateral nonlobar ground-glass opacities with septal thickening in 66.6% (N = 3, 2 out of 3); fibroparenchymal scarring in 33.3% (N = 3, 2 out of 3); subpleural fibrotic bands in the left lower lobe in 33.3% (N = 3, 2 out of 3), and pericardial effusion in 33.3% patients (N = 3, 2 out of 3). Dissemination of mucormycosis has been reported in 25% by Buil et al.[42] (N = 4, 1 out of 4) and 2.1% by Patel et al.[7] (n = 187, 4 out of 187) in their study population [Table 1].

Details of COVID-19 laboratory parameter

Very few studies reported laboratory parameters in the C-ROCM cases. Gangneux et al.[41] in their study population of invasive fungal infection reported that 64.2% (n = 57) patients presented lymphopenia at the time of first presentation. Bayram et al.[14] in their study population of C-ROCM patients reported that there was increased level of D-dimer (mean 1362.4 ± 468.9 μg/L, n = 11).

Microbiological profile of COVID-19–associated rhino-orbito-cerebral mucormycosis

Rhizopus species were the most common fungal species identified among positive fungal culture (7 studies,[27293136404243] prevalence 0.571, 95% C.I. 0.3–0.84). Mixed fungal growth was seen in 8.1% growth (6 studies,[6726384243] prevalence 0.081, 95% C.I. −0.007–0.169). Bacterial coinfection has been reported by 2 studies[2640] (prevalence 0.316, 95% C.I. 0.053–0.580).

Association of risk factor/co-morbidity for the occurrence of COVID-19–associated rhino-orbito-cerebral mucormycosis

Comorbidity

Diabetes mellitus

Diabetes is one of the most important comorbidities among C-ROCM patients. Pooled proportion of diabetes among C-ROCM patients were 0.79 (19 studies,[671425262728293132333435373940414243] 95% C.I. 0.686–0.895). Mean duration of diabetes at the time of diagnosis of C-ROCM was 5.9 year as reported by Sen et al.[26] (n = 6) while 4.4 years by Ravani et al.[30] (n = 31) and 12.1 years by Bayram et al.[14] n = 11). Mean HBA1C at time of presentation was varied as reported by different authors (7–15 mg/dl by Satish et al.,[24] n = 15; 6.5 mg/dl by Sharma et al.,[28] n = 25; 7.57 mg/dl by Ravani et al.,[30] n = 31; 9.7 mg/dl by Fouad et al.,[37] n = 12; and 10.2 gm/dl by Arjun et al.,[34] n = 12). Average fasting blood sugar at presentation is reported by single study[26] as 222.5 mg/dl (n = 6).

Other comorbidities

Apart from DM, other comorbidities that were prevalent among C-ROCM population were hypertension (12 studies,[1425262829343539404143] prevalence 0.435, 95% C.I. 0.340–0.530), chronic kidney disease (9 studies,[71425283034353743] prevalence 0.114, 95% C.I. 0.032–0.196), hematological malignancy (5 studies,[714373942] prevalence 0.066, 95% C.I. −0.012–0.144), coronary artery disease (3 studies,[142643] prevalence 0.235, 95% C.I. 0.052–0.419), acute renal failure (3 studies,[142843] prevalence 0.058, 95% C.I. −0.016–0.132) and ischemic heart disease (3 studies,[253537] prevalence 0.125, 95% C.I. −0.01–0.26).

Other rare comorbidities reported were atrial fibrillation (single study,[14] prevalence 9%, total population [n = 11]), chronic renal failure 36% (single study,[14] n = 11), hypothyroidism 10% (single study,[25] n = 10), hyperthyroidism 9% (single study,[14] n = 11), chronic liver disease 10% (single study,[25] n = 10), chronic obstructive pulmonary disease 9% (single study,[14] n = 11), and obesity 25% (single study,[42] n = 4).

Risk factors of C-ROCM

A significant proportion of C-ROCM patients in our study population received corticosteroid therapy as part of COVID-19 treatment protocol (proportion 0.857, 14 studies,[67142526272829323334394243] 95% C.I. 0.784–0.930). Mean duration of steroid use related to COVID-19 was reported by single study[14] (n = 11) as 8 days while another study[32] reported to be average 10–14 days (n = 25). Total corticosteroid dose equivalent received due to COVID-19 was reported to be 84 mg (cumulative dose of dexamethasone equivalent = 84 mg [18–1343 mg]) by single study[7] (n = 187).

Few studies reported use of mechanical ventilation as another risk factor. Pooled data showed that proportion of patients received mechanical ventilation as part of COVID-19 treatment protocol to be 0.454 (4 studies,[27293243] 95% C.I. −0.048–0.956). Similarly, use of supplemental oxygen has been reported as another risk factor by Mishra et al.[25] (40% patients, n = 10) and Arjun et al.[34] (80% patients, n = 10). Median worst PaO2/FiO2 for each patient was reported to be 124 (range 57–476) in the population as reported by Rabagliati et al. in their study population (N = 16, single study[40] ).

Arjun et al.[34] also reported use of broad-spectrum antibiotic as risk factor of occurrence of C-ROCM in their study population (90%, n = 10). No data were available regarding use of iron chelating in any of the included studies.

Intervention and outcomes analysis

Antifungal therapy

Amphotericin B was the mainstay of systematic as well as local antifungal therapy used in all studies. Pooled result showed that the proportion of C-ROCM patients receiving amphotericin B was 0.930 (14 studies,[671425262728293031323435394243] 95% C.I. 0.873–0.987) among which proportion of liposomal amphotericin B given was 0.902 (12 studies,[671425262728293031344243] 95% C.I. 0.825–0.978). Apart from intravenous form, proportion of patients receiving intraorbital amphotericin B was 0.506 (4 studies,[14262831] 95% C.I. −0.048–1.062), by intravitreal route in 54.5% of patients (in patients with with C-ROCM associated panophthalmitis, single study[14] ). Voriconazole (2 studies,[2643] proportion 0.206, 95% C.I. −0.054–0.467) and posaconazole were another antifungals used in the management of ROCM cases although very few studies reported.

Other concomitant therapeutic treatments given were broad-spectrum antibiotics (100% patients single study,[43] n = 3), tocilizumab (10% cases, 1 study,[25] n = 10), vasopressor (100% patients, single study,[43] n = 3) and intravenous (I.V.) dexamethasone (2 studies[1427] 100% patients, n = 21).

Surgical treatment

In C-ROCM cases, 70.9% cases were required mechanical debridement of para nasal sinus (13 studies,[14242526272829303234353743] proportion 0.709, 95% C.I. 0.587–0.832) while “functional endoscopic sinus surgery” was possible in 71.2% cases (4 studies,[6252627] proportion 0.712, 95% C.I. 0.416–1.007). Orbital decompression was done in 10% cases[25] (single study, n = 10) while exenteration was done in 21.2% patients (10 studies,[6242526272829313237] proportion 0.212, 95% C.I. 0.092–0.333). Other surgeries performed were maxillectomy in 34.6% patients (3 studies,[62527] n = 37, 95% C.I. 0.004–0.688), ethmoidectomy in 10% (single study,[25] n = 10), and craniotomy in 10% patients (single study,[25] n = 10).

Outcome analysis

Overall treatment success was seen in 60.8% patients [15 studies,[67142425272829343537394243] 95% C.I. 0.470–0.745, Figure 2a], while continued intracranial spread was seen despite treatment in 42.8% ROCM patients (13 studies,[6714242527282932353739] 95% C.I. 0.297–0.560). Treatment failure/mortality was seen in 34.4% of cases [15 studies,[6714242526272829343537394243] 95% C.I. 0.205–0.483, Figure 3a], with mean time to death found to be 19 days (3 studies,[293542] n = 16, 95% C.I. 3–35 days).

Figure 2

(a) Treatment success in C-ROCM population, (b) subgroup analysis of treatment success in C-ROCM patients based on severity of COVID-19

Figure 3

(a) Treatment failure/mortality in C-ROCM population, (b) subgroup analysis treatment failure/mortality based on severity of COVID-19

In the pooled result, high heterogeneity was seen while analyzing “treatment success” and “treatment failure” as indicated by I2 84% and 88% respectively [Figure 2a and Figure 3a]. To explore the etiology of “high heterogeneity,” we conducted “subgroup analysis” on the basis of “studies including “only severe and critical COVID-19 population” versus studies including “patients of all severity category of COVID-19” and study design (prospective vs. retrospective study). The high heterogeneity was not explained by subgroup analysis on the basis of COVID-19 severity [Figures 2b and 3b]. However, while doing subgroup analysis based on study design (prospective vs. retrospective study), very less heterogeneity was seen among prospective studies as compared to retrospective studies both in mortality and treatment success [Supplementary Figures 6 and 7].

Association between occurrence of C-ROCM and diabetic control

Prevalence of diabetes among C-ROCM cases was found to be 23.1% in controlled versus 68.5% in uncontrolled DM patients (11 studies,[61426262728293132374243] OR 0.150, 95% C.I. 0.041–0.544, I2 = 64.5%, random-effect model, P = 0.001), indicating significant association between uncontrolled DM with C-ROCM patients [Figure 4].

Figure 4

Association of controlled versus uncontrolled DM for occurrence of C-ROCM

Association between C-ROCM and COVID-19 severity

No significant difference was found with regard to occurrence of C-ROCM and COVID-19 severity (nonsevere vs. severe-to-critical COVID-19, 8 studies,[2425272829383942] OR 0.930, 95% C.I. 0.212–4.087, I2 = 74%, random-effect model, P = 0.923) [Supplementary Figure 8].

Meta regression

We evaluated the effect of different covariates on the incidence of occurrence of C-ROCM associated mortality in “univariate meta regression model.” Covariates that were reported in 10 or more studies were evaluated. The covariates evaluated were diabetes, age, male sex, liposomal amphotericin B treatment, and mechanical debridement. Among all these, only mechanical debridement has significant impact on mortality (i.e., protective effect as indicated by negative correlation, −0.004 P = 0.002) [Supplementary Table 3 and Supplementary Figure 9]. All other covariates did not show significant correlation with ROCM-associated mortality.

Supplementary Table 3

Univariate meta regression to evaluate the association between mortality and different covariates

Parameter	Number of studies	Intercept parameters		Covariate parameters					Overall model omnibus P
		Coefficient	P	Coefficient	Confidence interval		SE	P
					Lower	Upper
Age	13[2345,7,11,12,15,17,19,212223]	0.593	0.301	−0.005	−0.024	0.015	0.010	0.629	0.629
Male sex	13[234567,11,12,15,17,19,21,22]	0.251	0.363	0.001	−0.006	0.008	0.003	0.800	0.800
DM	14[234567,11,12,15,17,19,22,23]	0.068	0.858	0.003	−0.006	0.011	0.004	0.557	0.557
Liposomal amphotericin B therapy	10[234565,7,11,17,19,22,23]	0.727	0.144	−0.005	−0.016	0.006	0.005	0.368	0.368
Debridement	10[1234,6,7,12,15,17,22,23]	0.548	<0.001	−0.004	−0.007	−0.002	0.001	0.002	0.002**

*Statistically significant. SE=Standard error

Discussion

This systematic review and meta-analysis provides the global scenario of C-ROCM, its epidemiology, clinical presentation, and microbiological spectrum, which became a matter of concern due to sudden rise of mucormycosis/black fungus cases in the whole world parallel to ongoing COVID-19 pandemic.

Our meta-analysis showed that occurrence C-ROCM was more prevalent in elderly population with a mean age of presentation of 54.6 years and male gender being commonly affected. The lag period between COVID-19 diagnosis and diagnosis of ROCM was found to be average of 20 days (95% C.I. 15–24 days).

The most common presenting feature in C-ROCM was proptosis (60.6%), lid edema (60.7%), ophthalmoplegia (57.3%), loss of vision (53.7%), facial edema (34.7%), and ptosis (4 studies, 72.7%). Involvement of palate was not uncommon (35.6%). Rare occurrence of panophthalmitis (n = 4/7) and endophthalmitis[1435] (6/11) is also being reported as sight-threatening complication though in very few studies. Although intracranial dissemination was frequently seen owing to the contiguous spread, pulmonary involvement[7404243] (21/210) or dissemination to other organs was very rare[742] (5/191).

CT scan and MRI remain the most common radiological tool for diagnosis, evaluation of the extent of disease, and monitoring of response to treatment.[62628] Rhino-orbital involvement was the most common (70.8%) presentation followed by sinonasal (54.9%) and cerebral involvement (42.8%). Cavernous sinus involvement was seen in 40.7% patients. Ethmoid sinus was the single most common paranasal sinus involved[14283034] (80% by Arjun et al.[34] [n = 10], 90.9% by Bayram et al.[14] [n = 11], and 100% involvement reported by Sharma et al.[28] [n = 23]).

Rhizopus species were the most common species identified (57.1%) in total ROCM patients in our meta-analysis.

Diabetes was the most common comorbid condition in C-ROCM patients (79%). After diabetes, the most common associated comorbidities found among C-ROCM patients were hypertension (43.5%), coronary artery disease (23.5%), ischemic heart disease (12.5%), chronic kidney disease (11.4%), and hematological malignancy (6.6%).

Among other risk factors, 85.7% C-ROCM cases received corticosteroid; 12/20, i.e., 61%, received supplemental oxygen therapy;[2534] and 14/43, i.e., 45.4%, also received mechanical ventilation as part of COVID-19 therapy. One hypothesis says, increased use of corticosteroid as a part of COVID-19 treatment protocol served as exaggerating factor for glucose homeostasis which act as predisposing condition for the patients to opportunistic fungal infection.[6444546] However, the dose of corticosteroid used as part of COVID-19 treatment protocol is not reported in maximum of the study. Use of broad-spectrum antibiotic as the risk factor is also reported by single study.[34]

In our meta-analysis, significant association (P = 0.001) was found between uncontrolled DM (68.5%) with C-ROCM patients as compared to controlled diabetes (23.1%). However, no association was seen between occurrence of COVID-19 severity and C-ROCM. Similarly, no significant difference was seen with respect to occurrence of C-ROCM among concurrent COVID-19 population versus post-COVID-19 population (OR 0.306, 95% C.I. 0.042–2.246, P = 0.244).

Alteration of iron metabolism occurs in severe COVID-19[44] and iron overload (excessive free iron seen in acidosis) are other established risk factors for opportunistic fungal infection.[4748] However, none of the included in our meta-analysis has reported or gave specific consideration in this aspect.

Antifungal therapy was the mainstay of treatment both as presumptive and postdiagnosis therapy, and I.V. amphotericin B (93%) has been started as definitive treatment in all of the included studies in our meta-analysis, while liposomal amphotericin B was given in 90% of patients. Amphotericin B was given by intraorbital route in case of orbital involvement by few authors (16/42, 50.5%) and intravitreal route in case of panophthalmitis (6/11, 54.5%, single study[14] ); however, details of the antifungal susceptibility are not reported in maximum of the studies. The impact of liposomal amphotericin B needs further evaluation and newer antifungal option has to be taken into consideration.[495051525354555657585960]

Mechanical debridement was the commonly performed surgical procedure (70.9%) followed by maxillectomy (15/37, 34.6%), while exenteration was required in 21.2% patients. Less frequently performed surgeries were ethmoidectomy (1/10, i.e., 10%, single study[25] ), orbital decompression (1/10, i.e., 10%, single study[25] ), and craniotomy (1/10, i.e., 10%, single study[25] ). Although complete removal of infected tissue by surgical debridement at the earliest in addition to antifungal therapy is recommended as definitive treatment modality in patients with C-ROCM, it is not possible in all the cases, especially in patients with advanced disease at presentation or due to fitness issue for general anesthesia in maximum of the studies, leading to higher mortality rate.

Regarding the treatment outcome, treatment success was seen in 59.8% patients while continued intracranial spread was seen in 30.9% and treatment failure/mortality was seen in as high as 37.3% of cases despite treatment. While exploring the source of high heterogeneity, in treatment success and treatment failure, no difference was seen while doing subgroup analysis based on studies including only severe and critical versus studies including all category COVID-19 population. On the other hand, study design (prospective vs. retrospective study) seems to explain high heterogeneity in our meta-analysis. On univariate analysis for possible correlation of comorbidity and risk factors, mechanical debridement shows significant impact on mortality in C-ROCM patients (i.e., protective effect as indicated by coefficient − 0.004, P = 0.002). Again requirement of only debridement indirectly indicates a less severe form of disease without intracranial extension and without requiring other more invasive therapy. The protective effect of debridement may thus be explained in a dual manner.

Strength and weakness

This current meta-analysis provides the most recent and largest overview of the C-ROCM worldwide, which is the first meta-analysis published in this regard also. However, limited data regarding laboratory parameter of C-ROCM patients are available.

Conclusion

Our meta-analysis showed that occurrence C-ROCM was more prevalent in elderly population and among males with the most common presenting feature being loss of vision, proptosis, lid edema, ophthalmoplegia, facial edema, and ptosis. Although intracranial dissemination was frequently seen, pulmonary involvement or dissemination to other organs was very rare. CT scan and MRI remain the most common radiological tool for diagnosis and for evaluation of the extent of disease. KOH positivity for Mucor species was seen in 62% patients while 59.3% culture was positive for Mucor, Rhizopus species was the most common species to be identified (44.8%). 67.7% of the population had concomitant diabetes and 85.7% cases received corticosteroid as part of COVID-19 therapy. Compared to controlled DM, the incidence of ROCM was higher among patients with uncontrolled DM. However, no association was seen between COVID-19 severity and occurrence of ROCM. Amphotericin B, more specifically liposomal amphotericin B, was a mainstay of therapy along with other surgical management options. However, despite treatment, mortality/treatment failure was seen in 37.3% of cases.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Preferred Reporting Items for Systematic Review and Meta-Analysis flow diagram of included studies

Publication bias Funnel plot showing publication bias in case of mortality associated with COVID-19associated rhino-orbital-cerebral mucormycosis

Sensitivity analysis in terms of mean age of presentation by excluding Bayram et al. and Diwakar et al.

Mean duration between COVID-19 diagnosis and rhino-orbital-cerebral mucormycosis diagnosis

Comparison of incidence of C-ROCM between concurrent vs. post-COVID 19 patients

Mortality subgroup analysis prospective versus retrospective study design

Treatment success subgroup analysis prospective vs. retrospective study design

Occurrence of C-ROCM between non-severe vs. severe to critical COVID-19 population

Meta regression plot showing correlation of mechanical debridement with mortality due to C-ROCM