Emergence of Azole-Resistant Aspergillus fumigatus from Immunocompromised Hosts in India

This prospective study shows that the rate of azole-resistant Aspergillus fumigatus (ARAF) in an immunocompromised Indian patient population with invasive aspergillosis (IA) is low, 6/706 (0.8%). This low rate supports the continued use of voriconazole as the first line of treatment.

ABSTRACT

This prospective study shows that the rate of azole-resistant Aspergillus fumigatus (ARAF) in an immunocompromised Indian patient population with invasive aspergillosis (IA) is low, 6/706 (0.8%). This low rate supports the continued use of voriconazole as the first line of treatment. However, the ARAF isolates from India in this study exhibited three kinds of unreported cyp51A mutations, of which two were at hot spots, G54R and P216L, while one was at codon Y431C.

TEXT

Invasive aspergillosis (IA) is a disease of concern, as it is a leading cause of death in patients with hematological malignancies and in transplant recipients (1). Voriconazole is considered the drug of choice for primary therapy in IA cases (especially with cases of invasive pulmonary aspergillosis), while liposomal amphotericin B (L-AMB), caspofungin, and posaconazole are preferred as salvage therapy drugs (1). Antifungal drug resistance has mainly been reported among the triazoles (itraconazole, voriconazole, and posaconazole) (2–4).

In the resistant isolates, further studies of mutations provide a better understanding of their associations with particular drug susceptibility patterns. Aspergillus fumigatus is the most commonly isolated resistant species, and modifications of the target gene of triazoles, the cyp51A gene, have been found to be correlated with specific triazole resistance phenotypes (5–7). European data, especially that from the Netherlands, has added extensively to the subject (7–13). However, until this date, there are only two reports from India, wherein authors have mainly found TR34/L98H mutations in clinically isolated azole-resistant A. fumigatus (ARAF) strains (14, 15). While the world discusses ways to combat this deadly fungus (ARAF) that infects humans, the lack of data from India makes it difficult to comprehend its severity from the perspective of our country. Thus, this work is aimed at understanding the prevalence of ARAF and associated mutations in clinically suspected immunocompromised patients in India.

In this prospective study, a total of 1,416 azole-naive immunocompromised patients suspected of IA were enrolled over a period of 4 years (2012 to 2016). They were classified using the revised European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) diagnostic criteria, modified by including nonspecific radiological findings (16). There were 706/1,416 (49.8%) IA cases (8 proven IA and 698 probable IA) with 122/706 (17.3%) culture positivity and 6/706 (0.8%) ARAF, and there were 710/1,416 (50.2%) cases with no IA (444 possible IA and 46 [46/710; 6.5%] colonization cases with repeat isolation of same Aspergillus species in culture and direct examination negative for septate hyphae) and no ARAF. One probable reason for this high number of IA cases could have been a potential bias in selection of clinically suspected fungal cases. These patients had not responded to antibiotics for >72 h after admission. The isolates were identified following conventional mycological procedures, including direct specimen microscopy and growth on Sabouraud's dextrose agar (SDA) at a range of temperatures, namely, 25°C, 37°C, and 45°C (17, 18). Direct examination of specimens revealed that 128/1,416 (9%) tested positive for hyaline septate hyphae. However, the culture grew Aspergillus spp. in 122/1,416 (8.6%) clinically correlated cases (97 Aspergillus flavus, 22 Aspergillus fumigatus, two Aspergillus terreus and one Aspergillus niger) and were tested for their susceptibility profiles. For ARAF screening, all 32 A. fumigatus isolates (22 [68.7%] clinically correlated and 10 [37.5%] colonizers) were tested using the standard broth microdilution assays with Clinical and Laboratory Standards Institute (CLSI) approved standard M38-A2 and the European Committee on Antimicrobial Susceptibility Testing (EUCAST) (E.Def. 9.3) guidelines for molds (19, 20). For CLSI methodology, the proposed epidemiological cutoff values (ECVs) were followed, namely, itraconazole, 1 μg/ml; voriconazole, 1 μg/ml; posaconazole, 0.5 μg/ml; and amphotericin B, 4 μg/ml. Similar cutoffs were followed for micafungin as that reported for caspofungin, 0.25 μg/ml (21, 22). For EUCAST methodology, the breakpoints were followed as per the clinical breakpoints (version 8.1) published on EUCAST website, namely, for amphotericin B, itraconazole, and voriconazole, ≥2 μg/ml, and for the remaining drugs, i.e., posaconazole, caspofungin, and micafungin, ≥0.25 μg/ml (23). Detailed statistical analysis showed a high recorded agreement (94 to 98%) between the two methodologies. However, intraclass coefficients (ICCs) were found to be 0.97 to 0.98 for the azoles and polyene tested but were found to be poor for the echinocandins tested (Table 1).

TABLE 1.

Susceptibilities of Aspergillus fumigatus to various antifungals and concordance and intraclass coefficients between CLSI M38-A2 and EUCAST guidelines

Antifungal agents	CLSI (%) (n = 32)		EUCAST (%) (n = 32)		Concordance (%)	ICC (95% CI)a
	Susceptible (no. [%])	Resistant (no. [%])	Wild type (no. [%])	Non-wild type (no. [%])
Itraconazole	26 (81.25)	6 (18.75)	26 (81.25)	6 (18.75)	98	0.98 (0.86–0.99)
Voriconazole	31 (96.87)	1 (3.12)	31 (96.87)	1 (3.12)	97	0.98 (0.93–0.99)
Posaconazole	31 (96.87)	1 (3.12)	31 (96.87)	1 (3.12)	94	0.96 (0.91–0.98)
Amphotericin B	31 (96.87)	1 (3.12)	31 (96.87)	1 (3.12)	97	0.97 (0.90–0.99)
Caspofungin	32 (100)	0	32 (100)	0	56	0.65 (0.30–0.83)
Micafungin	32 (100)	0	32 (100)	0	56	0.65 (0.30–0.83)

^aICC, intraclass coefficient; CI, confidence interval.

In the study, in vitro antifungal resistance (irrespective of class of drugs) was noted in 2.3% (16/706) of IA patients, in 13.1% (16/122) of IA isolates and in 2.2% (5/229) of voriconazole-exposed IA patients (including 191 voriconazole recipients and 113 voriconazole-amphotericin B recipients) (Table 2). This high rate of resistance is in accordance with previously published literature (24, 25). It is worth noting that a much higher rate has also been previously reported (13, 26).

TABLE 2.

Proportion of in vitro antifungal resistance with regard to treatment given in this study

Treatment	All patients		IA patients		IA isolation cases		Susceptible IA isolation cases		Resistant IA isolation cases		A. fumigatus isolation cases		Susceptible A. fumigatus isolation cases		Resistant A. fumigatus isolation cases
	Total (no. [%])	Poor outcome (no./total, [%])	Total (no. [%])	Poor outcome (no./total, [%])	Total (no. [%])	Poor outcome (no./total, [%])	Total (no. [%])	Poor outcome (no./total, [%])	Total (no. [%])	Poor outcome (no./total, [%])	Total (no. [%])	Poor outcome (no./total, [%])	Total (no. [%])	Poor outcome (no./total, [%])	Total (no. [%])	Poor outcome (no./total, [%])
Any	1,416 (100)	192/1,416 (13.6)	706 (100)	107/706 (15.2)	122 (100)	44/122 (36)	106 (100)	32/106 (30.2)	16 (100)	12/16 (75)	32 (100)	17/32 (53)	26 (100)	9/26 (34.6)	6 (100)	4/6 (66.7)
None	440 (31)	68/440 (15.5)	147 (20.8)	25/147 (17)	0	0	0	0	0	0	0	0	0	0	0	0
Voriconazole	191 (13.5)	12/191 (6.3)	116 (16.4)	9/116 (7.8)	25 (20.5)	2/25 (8)	23 (21.7)	2/23 (8.7)	2 (12.5)	0	9 (28.1)	0	8 (30.8)	0	1 (16.7)	0
Amphotericin B	352 (24.9)	59/352 (16.8)	183 (25.9)	36/183 (19.7)	36 (29.5)	22/36 (61.1)	28 (26.4)	15/28 (53.4)	8 (50)	7/8 (88)	9 (28.1)	8/9 (88.9)	7 (27)	6/7 (85.7)	2 (33.3)	2/2 (100)
Caspofungin	25 (1.8)	5/25 (20)	13 (1.8)	2/13 (15.4)	5 (4)	2/5 (40)	5 (4.7)	2/5 (40)	0	0	1 (3.1)	0	1 (3.8)	1/1 (100)	0	0
Amphotericin B and voriconazole	162 (11.4)	18/162 (11.1)	113 (16)	14/113 (12.4)	27 (22.1)	7/27 (26)	24 (22.6)	5/24 (20.8)	3 (18.8)	2/8 (66.7)	6 (18.8)	5/6 (83.8)	4 (15.4)	0	2 (33.3)	1/2 (50)
Other drug combinations	246 (17.4)	30/246 (12.2)	134 (19)	21/134 (15.7)	29 (23.8)	11/29 (38)	26 (24.5)	8/26 (30.8)	3 (18.8)	3/3 (100)	7 (1.9)	4/7 (57.1)	6 (23)	2/6 (33.3)	1 (16.7)	1/1 (100)

Globally, there has been an alarming increase of azole resistance in A. fumigatus (2, 4, 27). From Manchester, a 2% rate of ARAF was reported in 2000 (3) whereas from 2007 to 2008 this rate increased to 15%, and later in 2009 to 20% (24, 25, 27). Similar increases were reported from the Netherlands, namely, from 2.5% in 2000 to 4.9% in 2002 and 6.6% in 2004 to 10% in 2009 (24). Azole resistance had been explored in multicenter studies, including the ARTEMIS global surveillance study, which showed a 5.8% rate of ARAF (28). In another study by the Surveillance Collaboration on Aspergillus Resistance in Europe (SCARE) network, the prevalence rate of ARAF was 3.2% (29). In this study, ARAF was seen in 6/706 (0.8%) IA cases, which validated voriconazole as the first line of drug treatment for IA in our hospital settings. Similar to our study, findings from a single center in Paris, France, found ARAF in only 1/152 (0.7%) IA cases from immunocompromised patients (30). Previously, the same French center isolated ARAF at a 1.1% rate from a specific hematological malignancy patient group (10). Geographically, a wide range of triazole resistance rates has been reported in A. fumigatus clinical isolates, including <1 to 3.6% in the United States (31, 32), 4% in China (33), 4.5% in Kuwait (34), 11% in Japan (35), and 2% in India (36). Drug-resistant IA has been associated with a high mortality rate (4, 5, 26, 37, 38).

In this study, the overall observed 30-day mortality rate was 13.6% (192/1,416) and was much lower than what has been reported in the literature (mostly ≥50%) (39–42). This may be due to reporting of a lesser time to mortality (∼1 month), whereas in most of the studies time to mortality is ≥3 months. Poor prognosis has been previously correlated with varied mycological diagnostic criteria, including culture, galactomannan antigen (GM), and PCR positivity (43–47). It has been associated with persistent higher GM values, and it was recently found to be linked with higher PCR fungal load (47).

However, among the parameters observed in this study, the 30-day outcome was notably poor in both culture-positive (44/122; 36.1%) versus culture-negative IA cases (63/584; 10.8%) and resistant versus susceptible isolation cases (Table 2). Similarly, 66.7% (4/6) of ARAF isolation cases had poor outcomes compared to 34.6% (9/26) of azole-susceptible A. fumigatus isolation cases (Table 2).

Indian data on IA mortality varies between different patient groups. In a study on acute leukemia and hematopoietic stem cell transplantation (HSCT) patients, the overall observed mortality rate was 15.3% (all causes), and according to the EORTC/MSG classification of IA cases, the mortality was 36% in possible IA, 41.1% in probable IA, and 100% in proven IA cases (48). Recently, in hematological malignancy patients, the 4-month observed mortality rate due to IA was found to be 30% (49).

The mechanisms of azole resistance in the species have been extensively studied in the Western world, and the results are alarming and call for regular surveillance worldwide. However, the only Indian clinical data are from one center, wherein a total of 14 ARAF cases, primarily having TR34/L98H mutations in their cyp51A gene, have been reported from two different studies (14, 15). The same group pioneered resistance surveillance of environmental isolates and reported major TR/L98H mutations in 44/630 ARAF isolates (42); they also reported the first TR46/Y121F/T289A mutations in 6/1,210 resistant environmental isolates (50). TR/L98H is the most common cyp51A mutation and has been previously reported in this patient population from various parts of the world, namely, Nijmegen, Netherlands (Europe) (7, 9, 12, 13); Germany (Europe) (11); Poland (Europe) (51), the United States (North America) (52); Taiwan (53); Australia (54); the Middle East (34, 55); and New Delhi, India (15). However, this mutation has been linked with the environmental spread of resistance due to the increasing use of fungicidal azoles in agricultural practices in Copenhagen, Denmark (Europe) (56); Bogotá, Colombia (Latin America) (57); Kuwait (Middle East) (34); and New Delhi, India (36).

In this study, cyp51A mutations were sought in all A. fumigatus isolates using the forward primer P450A (58) and reverse primer CYP3 R (59), yielding a PCR product of ∼1,500 bp. All of the six A. fumigatus isolates showing elevated MICs harbored mutations in the cyp51A gene, leading to azole resistance (Table 3).

TABLE 3.

Characteristics of patients harboring azole-resistant A. fumigatus

Patient ID	Sex, age (yrs)a	Date of admission	Underlying diseaseb	Clinical sample	EORTC/MSG criterion	ΔGMd	cyp51A mutatione	GenBank accession no.	MIC (μg/ml) for:			Resistant antifungal(s)f	Antifungal treatment(s)g	30-day outcome
									ITR	VOR	POS
1	M, 18	16 February 2013	ALL	BALc fluid	Probable	−1.52	P216L	MF148150	8	0.5	0.06	ITR	AMB	Died
2	M, 54	19 April 2014	TB	Tissue	Proven	−0.4	G54R	MF148151	4	4	0.06	ITR, VOR	AMB, VOR	Died
3	F, 18	7 March 2013	ALL	BAL fluid	Probable	−2.78	Y431C	MF148152	16	0.5	2	ITR, POS	AMB, VOR	Alive
4	M, 5.2	21 April 2013	ALL	BAL fluid	Probable	−0.802	P216L	MF148153	8	0.25	0.06	ITR	AMB	Died
5	M, 6	19 September 2013	TB	BAL fluid	Probable	−1.3	Y431C	MF148154	16	1	0.06	ITR	AMB, CAS	Died
6	F, 5	11 May 2014	CML	Sputum	Probable	−1.199	P216L	MF148155	8	0.25	0.125	ITR	VOR	Alive

^aM, male; F, female.

^bALL, acute lymphoid leukemia; TB, tuberculosis; CML, chronic myeloid leukemia.

^cBAL, bronchoalveolar lavage.

^dΔGM = GM1 − GM2; GM, galactomannan antigen.

^eP, proline; L, leucine; G, glycine; R, arginine; Y, tyrosine; C, cysteine.

^fITR, itraconazole; VOR, voriconazole; POS, posaconazole.

^gAMB, ampshotericin B; CAS, caspofungin.

These were azole-naive patients at the time of sampling, which suggested a possible role for environmental transmission of resistance. India being an agrarian economy, the probable role of fungicides in spreading azole resistance cannot be ruled out (24, 60). Irrespective of previous reports from azole-naive patients, none of the three kinds of mutations reported in the study have been previously linked with environmental mechanisms of resistance spread. These mutations have not yet been reported from India. However, at codon 54, a substitution to glutamic acid (G54E) had been previously reported from an environmental isolate (36) and a clinical Indian strain (15). The nonsynonymous hot spot mutation (G54R) was only seen in one isolate, whereas the common mutation observed in a conserved region, P216L, was seen in three isolates. Both of these mechanisms prevented the docking of drug molecules and thereby produced the resistant strains (7, 61). The substitution to cysteine at codon 431 has been associated with a complex resistance mechanism involving upregulation/increased expression of cyp51A (62). The prevalence of similar resistance mechanisms (G54R, Y431C, and P216L) in the clinical and environmental isolates in India should therefore be investigated by further studies.

Recently, experts had discussed the treatment options in clinical cases from regions where ARAF strains were rampant (>10%) or moderate (5 to 10%). Even though they had a difference of opinion, they agreed on the use of national epidemiology and local susceptibility data to guide the choice of treatment (63).

Due to limited resources, we could not perform short tandem repeats of Aspergillus fumigatus (STRAf) typing to compare our isolates with isolates from worldwide. Another limitation of the study was the small number of biopsy and bronchoalveolar lavage samples.

In summary, this study supports the fact that azole resistant A. fumigatus is prevalent in India, which is an upsetting finding in a country where the climate supports fungal growth and antifungal drugs are expensive and mostly unaffordable. However, the low rate of ARAF in IA cases (0.8%) does not cause alarm about voriconazole as the first line of treatment. Our findings, combined with those of previous studies, emphasize antifungal susceptibility-directed treatment of IA for better management of patients and to avoid unnecessary use of toxic and expensive drugs.