Microbiology: Mould contamination in pharmaceutical drug products and medical devices

This review article will examine microbial testing conducted in the pharmaceutical industry directed towards the detection, enumeration and the identification of fungi, the clinical aspects of fungal contamination and the product recalls associated with fungal contamination. It is the belief of the author that microbiologists working in the pharmaceutical and medical device industries should give fungi more attention in their microbial contamination control programmes.

Compendial microbial tests

Traditional microbial testing methods described in the compendia, as referee tests, rely on the growth of microorganisms in culture media for detection, enumeration and selective isolation (USP <51>, <61>, <62> and <71>). These traditional methods continue to be used because of their long history of use, simplicity, effectiveness and suitability for use in all microbiological testing laboratories. These methods when cited in regulatory filings that have microbial specifications become product release and shelf-life tests using the recommended microbiological standards in USP <1111> for microbial enumeration and absence of specified microorganisms test (Table 1). However, the drivers of microbial testing should be the critical microbiological quality attributes associated with a specific drug product and the risk assessment of the potential for microbial contamination of that drug product.

Table 1: The Harmonized Recommended Microbiological Quality Requirement for Non-sterile Drug Dosage Forms

The USP compendial microbial chapters related to non-sterile drug products are:

• USP <51> Antimicrobial Effectiveness Testing
• USP <61> Microbiological Examination of Non-sterile Products: Microbial Enumerations Tests (Harmonized)
• USP <62> Microbiological Examination of Non-sterile Products: Tests for Specified Microorganisms (Harmonized)
• USP <71> Sterility Tests
• USP <1111> Microbiological Quality of Non-sterile Pharmaceutical Products (Harmonized)
• USP <1112> Microbiological Attributes of Non-Sterile Pharmaceutical Products – Application of Water Activity Determinations
• USP <1113> Microbiological Characterization, Identification and Strain Typing
• USP <1117> Microbiological Best Laboratory Practices
• USP <1235> Water for Pharmaceutical Purposes

USP chapters below <1000> are termed general test chapters and those greater than <1000> are general informational chapters. The application of USP <61> and <62> as release tests to non-sterile drug products will be more fully discussed later in the review article. The minimal microbial acceptance requirements by dosage form are found in USP <1111> (See Table 1).

As highlighted in Table 1, the microbial enumeration requirements for fungi are more stringent than bacteria. Is it justified having the total combined yeast and mould count limits one log lower than the total aerobic microbial count limit? In terms of potential to cause human infection, bacterial pathogens are more numerous and significant than fungal pathogens. Fungal infections are usually associated with immune-compromised patients, however, infection-causing fungi may be more difficult to detect and identify, fungi may grow in drug products with lower water activities than bacteria, the therapeutic options are more limited and the mortality rates with immuno-compromised patients may exceed 40 per cent.

There are two parts to the compendial microbial tests; namely, the microbial enumeration tests (USP <61>) and the absence of specified microorganism tests (USP <62>). The microbial enumeration tests are subdivided into the Total Aerobic Microbial Count (TAMC), Total Combined Yeast and Mould Count (TCYMC), Most-Probable-Number (MPN) Count and quantitative test for bile-tolerant gram-negative bacteria.

As the TAMC uses a general microbiological growth medium Soybean-Casein Digest Agar (SCDA) that is capable of supporting the growth of both aerobic bacteria and fungi, P. aeruginosa ATCC 9027, S. aureus ATCC 6538, B. subtilis ATCC 6633, C. albicans ATCC 10231 and A. brasiliensis ATCC 16404 are the bacteria and fungi specified for growth promotion testing. C. albicans ATCC 10231 and A. brasiliensis ATCC 16404 are the fungi specified for growth promotion testing Sabouraud Dextrose Agar (SDA) used for TCYMC. Operationally, any colony growing on either a Soybean-Casein Digest Agar or a Sabouraud Dextrose Agar plate will be counted irrespectively if it is a bacterial, yeast or mould colony. This suggests that in some circumstances based on testing history it may not be necessary to conduct counts on both media when low numbers of faster growing bacterial colonies are not expected to crowd out fungal colonies. Sabouraud Dextrose Agar was introduced by the pioneering French mycologist Sabouraud for the selective cultivation of dermatophytes such as Trichophyton mentagrophytes in the presence of high numbers of skin bacteria. The high dextrose content and low pH is favourable for the growth of fungi, especially dermatophytes, and is inhibitory to contaminating bacteria that may be found in clinical specimens that could overgrow the plate. Antibiotics such as chloramphenicol, at a concentration of 0.04 per cent, may be added to further suppress the growth of bacteria on the plates. This could be adopted for routine drug product monitoring where the recommended microbiological quality criterion may be TAMC not more than 10² cfu/g and TCYMC not more than 10¹ cfu/g. and more than one colony on a Sabouraud Dextrose Agar plate may result in a count exceeding the fungal criterion.

Is it truly possible to enumerate fungi? Most moulds are filamentous with aerial hyphae bearing spores. This mode of growth makes it questionable that a so-called colony-forming unit is truly representative of the fungal biomass within a product.

In addition, P. aeruginosa ATCC 9027, S. aureus ATCC 6538, and B. subtilis ATCC 6633 are the microorganisms specified for growth promotion testing Soybean-Casein Digest Broth for TAMC using the MPN-multiple tube method. The compendial chapter points out that this method is less accurate and precise than the membrane filtration and plate count methods and unreliable results may be obtained in the enumeration of mould.

Tests for the Absence of Specified Microorganisms that are applied to different dosage forms are found in USP <62> and <1111>. The screening tests are absence and quantitative of bile salt-tolerant gram-negative bacteria, absence of E. coli, P. aeruginosa, S. aureus, C. albicans, and Clostridium spp. (in one gram) and absence of Salmonella spp.(in 10 grams).

These specific microorganism tests generally consist of three steps: 1) general enrichment in soybean – casein digest or Sabouraud dextrose broth to increase the number of microorganisms, 2) selective enrichment using specialised broth and incubation conditions to select for the target microorganism, and/or 3) growth on solid diagnostic media for presumptive identification of the microorganisms. This presumptive identification is based solely on growth on the diagnostic media and/or colony colour and morphology, cellular morphology, Gram reaction and other diagnostic tests. The identity is confirmed using standard microbial identification tests. It may be possible, if justified, to eliminate the selective enrichment then you are routinely testing pharmaceutical products with a low bio burden to maximise the recovery of the widest range of microorganisms.

The growth-promotion microorganisms of the non-selective enrichment broth are S. aureus ATCC 6538, P. aeruginosa ATCC 9027, B. subtilis ATCC 6633 for soybean-casein digest broth and C. albicans ATCC 10231 and A. brasiliensis ATCC 16404 for Sabouraud Dextrose Agar or Broth.

Test for the absence of C. albicans

The test consists of selective enrichment in SD broth at 30 – 35°C for three to five days and subculture on SD agar at 30 – 35°C for 24 – 48 hours. Greater than ambient temperature, low pH and high dextrose concentration of the media favour pathogenic yeast. The addition of chloramphenicol to the SDA is inhibitory to a wide range of gram-negative and gram-positive bacteria. The growth of white colonies on SDA is indicative of the yeast C. albicans. The objectionable organism isolation rating is good as the schema is selective for yeast and mould especially pathogenic yeast due to the above ambient incubation temperature.

The production of pseudohyphae, germ tube and/or chlamydospore formation is indicative of C. albicans. The identity of any colonies isolated on the diagnostic media would be determined using the microbial identification method employed in your laboratory.

Sterility tests

In 1936, the sterility test was introduced for the testing of liquid products in USP XI. The original test method was seven day incubation at 37°C using a broth containing beef extract, peptone, sodium chloride, and dextrose. By USP XVII (1965), the sterility test had evolved to the use of Fluid Thioglycollate Medium incubated at 30 to 35°C for at least seven days, and Sabouraud Dextrose Medium incubated at 20 to 25°C for at least 10 days. The use of thioglycollate was a significant advance for the detection of anaerobes and the neutralisation of mercuricals used as preservatives in biological products. The use of Sabouraud Dextrose Medium assisted with the detection of yeasts and moulds. In 1970, Soybean–Casein Digest Medium was substituted for Sabouraud Dextrose Medium and the incubation for aseptically filled products was extended to 14 days, with a provision for seven day incubation for products terminally sterilised with a moist heat sterilisation process. Subsequently these exceptions for incubated time were wisely eliminated. Furthermore, the incubation time could be reduced from 14 to seven days when a membrane filtration method was used. The rationale for these changes is that it is inappropriate to use a medium such as Sabouraud Dextrose Medium in the sterility test when it does not support the growth of many bacteria. Additionally, membrane filtration is an improvement over direct inoculation because of the improved efficiency of removal of antimicrobial agents from the test specimen, the use of the total volume of product in each container instead of an aliquot from each container, and the concentration of the contaminating microorganisms within the broth culture.

The compendial sterility media and incubation conditions are a compromise for the detection of viable microorganisms in sterile products. In general, moulds require aerobic incubation using carbohydrate rich media, lower pH conditions and near ambient incubation temperatures. The expectancy is that fungi are more likely to be isolated in the soybean-casein digest broth incubated at 20 – 25°C for at least 14 days than in fluid thioglycollate broth incubated at 30-35°C.

Antimicrobial effectiveness testing

The effectiveness of an antimicrobial preservative system in a multiple-dose, aqueous drug product is demonstrated using the methods found in USP <51>. The difference in the antimicrobial effectiveness acceptance criteria for the log reduction of bacteria versus fungi for different dosage forms (Table 2) is a concession to the reduced effectiveness of antimicrobial preservatives against yeast and mould. The reader will notice that the acceptance criteria are not harmonised with the Ph. Eur. requiring log reductions for fungi. The author questions the less stringent requirements for fungi in USP <51> is justified.

Clinical aspects of fungal infections

According to the CDC website, fungal diseases are an increasing threat to public health due to the increase in both opportunistic and community-acquired fungal infections (www.cdc.gov/fungal). Opportunistic infections such as cryptococcosis and aspergillosis are becoming increasingly problematic as the number of people with weakened immune systems rises. These people include cancer patients, transplant recipients, surgical patients and people with HIV/AIDS. Hospital-associated infections such as candidemia are the fifth leading cause of bloodstream infections in the United States. Changes in healthcare practices can provide opportunities for new and drug-resistant fungi to emerge in hospital settings. Community-acquired infections such as coccidioidomycosis (Valley Fever), blastomycosis and histoplasmosis, are caused by fungi that are abundant in the environment. These types of fungi live in the soil, on plants or in compost heaps, and are endemic (native and common) throughout much of the US. Mycologists believe that climate change may be affecting these fungi, as even small changes in temperature or moisture can affect their growth.

Fungi cause a wide variety of diseases in humans, and the areas we discuss are listed below. Please also refer to the Infectious Disease Society of America-Mycoses Study Group (IDSA-MSG) Practice Guidelines for treating invasive mycoses. These cover aspergillosis, blastomycosis, candidiasis, coccidiodomycosis, cryptococcosis, histoplasmosis, and sporotrichosis and are available at the ISDA website (www.idsociety.org/pg/toc.html).

The prevalence of fungi in drug product recalls

A recent US survey of 144 recalls, for the seven year period from 2004 through 2011 of non-sterile branded pharmaceutical drug products (five per cent), over-the-counter drug products (42 per cent), cosmetics (31 per cent), medical devices (14 per cent) and dietary supplements (eight per cent of the total recalls) for microbiologically-related issues highlighted that the majority of these recalls (72 per cent) were associated with objectionable microorganisms⁴. The most frequently cited microorganisms in the recalls were the Burkholderia cepacia (34 occurrences), unspecified fungal contamination (19 occurrences), Bacillus cereus (nine occurrences), Pseudomonas aeruginosa (six occurrences), Elizabethkingia meningoseptica (five occurrences), Enterobacter gergovia (five occurrences), Pseudomonas putida (three occurrences), Pseudomonas spp. (two occurrences) and Salmonella spp. (two occurrences). Seventeen other occurrences were associated with a single species. Fungal contamination is the second most frequent cause of product recall. The unspecified fungal identification may imply multiple fungal contaminants or the inability of the manufacturers to identify the implicated fungi. As stated in the introduction, the author believes that the pharmaceutical industry is doing a poor job in area of mycology.

Case histories

The following case histories illustrate the involvement of mould contamination of the common pharmaceutical ingredient dicalcium diphosphate, aseptic processing areas subject to water damage, contact lens solutions, a compressed tablet, a skin-moisturising lotion and sterile compounded injectable steroid. It is hoped that these case histories will be informative to the reader.

Table 2: The Differences in the Antimicrobial Effectiveness Acceptance Criteria by Product Category for Bacteria and Fungi

Commonly used pharmaceutical ingredient

The presence of fungal contaminants in a pharmaceutical ingredient may have both health and regulatory consequences. On 21 September 2001, Pharmacia-Upjohn recalled 42 lots of their generic Glyburide tablets and three lots of their branded MICONASE tablets (Glyburide tablets, USP) for fungal contamination. The source of the contamination was traced to the tablet filler, Dibasic Calcium Phosphate Dihyrate, USP used in the formulation. In the 18 September 2001 warning letter that triggered the recalls, the firm was cited for inadequately investigating the sources of the fungal contaminants, failure to identify other Glyburide tablet lots manufactured with common excipient lots, failure to appropriately sample and test the excipient and failure to issue Field Alerts as required by GMP regulations. Subsequently it was determined that process air used to dry the excipient was contaminated by seasonal fungal spores, resulting in heterogeneous contamination of the excipient.

Aside from the GMP violations, why did the FDA issue a warning letter highlighting fungal contamination of a compressed tablet and multiple lots were recalled? Glyburide tablets are prescribed to control diabetes so the drug product is targeted for a specific patient population. According to a recent review article⁵, the common infections that can cause serious complications in patients with diabetes include community-acquired pneumonia, urinary tract infections, necrotising fasciitis and foot ulcers. In contrast, certain rare infectious diseases occur predominantly in patients with diabetes, including malignant otitis externa, rhinocerebral mucormycosis, emphysematous cholecystitis and emphysematous pyelonephritis. Because these conditions can have serious complications, progress rapidly and/or be life-threatening, prompt recognition and treatment is imperative. Glycemic control, screening for foot ulcers and educating patients about foot care, and vaccination against pneumococcus and influenza are critically important strategies for preventing infection in diabetes patients. It is unlikely that the fungal contaminants would expose diabetes to infections.

Mould contamination of sterile product manufacturing plants

Two examples of mould contamination of sterile manufacturing plants are contained in recent FDA warning letters. In October 2011, Sanofi Pasteur’s sterile product manufacturing facility in Toronto, Canada experienced flooding that led to water damage. Due to adverse trends in fungal isolation during environmental monitoring in Building 86 and questions from the Australian health authorities as to the state of validation of their sterility test for the BCG tuberculosis vaccine that resulted in the recall of four vaccine lots and multiple 483 observations arising from an April 2012 FDA inspection, the firm decided in July 2012 to halt production and repair the building. As documented in the 12 July 2012 FDA warning letter for a period from August 2010 through April 2012, there was no fewer than 58 non-conforming mould isolations without adequate investigation and corrective action. Also, the FDA cited an insufficient frequency of monitoring in relation to the duration of the fill, poor aseptic technique in the aseptic processing areas, and inadequacy of the firm’s disinfectant / sporicidal agent effectiveness studies with respect to fungal spores, and poor facility maintenance. This contributed to a worldwide shortage of the BCG vaccine.

An industry consensus related to the rules for trending and tracking mould isolates in aseptic processing areas and the frequency of mould identification in different classified areas is lacking.

Table 3: Major Fungi Associated with Human Infectious Disease

Contact lens solution

On 13 April 2006, the American medical device company Bausch and Lomb withdrew their contact lens solution ReNu with MoistureLoc^® from the US market due to fungal infection experienced by contact lens users. As of 30 June 2006, the US Federal Center for Disease Control and Prevention (CDC) identified 164 confirmed cases of the corneal infection, Fusarium keratitis, from 33 states and one US territory. Corneal transplants were required or planned in 55 cases. In 154 confirmed cases, individuals wore soft contact lens. The investigation demonstrated that the infections were significantly more likely to be associated with the use of ReNu with MoistureLoc^® than other lens solutions. Since the fungus Fusarium was not found at the manufacturing facility, in solution filtrate, or unopened solution bottles, and the implicated lots were not clusters in time, it indicated that the microbial contamination was extrinsic and not intrinsic. This was confirmed by the finding that amongst 39 Fusarium isolates identified, at least 10 different Fusarium species were detected, comprising 19 unique genotypes using nucleic acid sequencing. If the fungal contamination was intrinsic, the isolates would a single or limit number of species and genotypes.

Fusarium is a filamentous fungus commonly found in soil as a plant pathogen and occasionally in water systems. It is a major cause of fungal keratitis in tropical and subtropical areas often associated with eye damage. Fungal keratitis is relatively rare amongst contact lens wearers, comprising less than five per cent of microbial keratitis. In the majority of the cases in this outbreak, individuals were using lens made of non-silicone hydrogel, high-water content non-ionic polymers and ReNu MoistureLoc^® lens solution. Based on a case-control study conducted by the CDC, the reuse of solution in the contact lens during storage in the lens case appeared to contribute to the incidence of infection. Fusarium was most frequently isolated from used contact lens cases. The pattern of Fusarium isolation, speciation and genotyping favours a cause of extrinsic contamination of solution bottles or lens cases from the local user environment. Whether Fusarium solani or F. oxysporum was used as a challenge organism in the Antimicrobial Effectiveness Test against pelagic or biofilm cultures during product development is unknown to the author.

Contrary to the 22 August 2006 statement by Bausch and Lomb, who cited widespread non-compliance with recommended contact lens care and cleaning practices as the cause, the CDC concluded hygiene practices were not the major cause of the outbreak as a similar pattern was seen among case patients and control. An exception was storing lens by reusing solution already in the lens case. CDC believes that interaction between MoistureLoc^® solutions, Fusarium, and possibly the lens case or contact lens was a causal factor. As the CDC investigation revealed the majority of the keratitis cases individuals were using lens made of non-silicone hydrogel, high-water content non-ionic polymers and ReNu with MoistureLoc^® lens solution. Based on a case-control study conducted by the CDC, the reuse of solution in the contact lens during storage in the lens case appeared to contribute to the incidence of infection. The MoistureLoc^® formulation contains two ingredients not contained in other marketed lens solutions. They are a moisture-retaining polysaccharide Polyquarterium 10 that holds water on the surface of the lens surface (hence the name MoistureLoc^®) and the preservative Alexidine (a cationic bisbiquanide). Also, Poloxamer 407, a non-ionic surfactant, helps retain moisture on the lens surface, remove debris, and retain protein in their native state. The author believes that the formulation encouraged fungal biofilm formation, preventing the action of the preservative system. Fusarium was most frequently isolated from used contact lens cases. As stated earlier, the pattern of Fusarium isolation, speciation and genotyping favoured a cause of extrinsic contamination of solution bottles or lens cases from the local user environment.

Allopurinol tablets

On 9 March 2009, the Hong Kong Board of Health reported that four batches of Allopurinol tablets manufactured by a local pharmaceutical company Europharm were found to be grossly contaminated with the fungus Rhizopus microsporus (>10³cfu/g). At least five patients at Queen Mary’s Hospital receiving immunosuppressive therapy and treated for the side effect hyperuricemia contracted intestinal mucomycosis and died⁶.

Zygomycosis is the third most common invasive fungal infection after candidiasis and aspergillosis. The incidence of zygomycosis is rising due to multiple factors including the increasing use of potent immunosuppression, stem cell and organ transplants and possibly selection for Zygomycetes by prior treatment with broad-spectrum antifungal therapy, which has no activity against Zygomycetes. Rhizopus species accounted for 218 of 465 (47 per cent) patients with zygomycosis but R. microsporus has only been isolated from 11 of 465 (two per cent) patients. Triazoles, such as itraconazole, fluconazole and voriconazole, have little activity against Zygomycetes and therefore are an ineffective treatment when zygomycosis is suspected. However, posaconazole, a new broad-spectrum triazole, has proven successful as salvage therapy in treatment of breakthrough zygomycosis.

The tablets were manufactured at the EuroPharm Hong Kong facility using a wet granulation that was dried in a tray dryer oven at 50°C for four hours to a water content of three per cent. The granulation was held at 20°C for five to 14 days prior to tablet compression. A typical formulation is allopurinol, 100 or 300 milligrams, corn starch, FD&C Yellow No. 6 Lake (yellow tablets only), lactose, magnesium stearate and povidone. A possible source of the Rhizopus microsporus was the corn starch used in the tablet manufacture that contained two cfu of Rhizopus/g. Although the ascospores of Rhizopus microsporus are thermotolerant and would survive four hours at 50°C, it appears unlikely that a granulation dried to three per cent water content and stored at 20°C for five to 14 days prior to tablet compression would become highly contaminated due to low water activity and the short time frame.

According to the Handbook of Pharmaceutical Excipients, corn starch, NF has a water content of 11 – 14 per cent (water activity 0.6 to 0.9). The corn starch in a granulation dried to three per cent has an estimated water activity of 0.1 to 0.2 and would not support the growth of Rhizopus microsporus if stored at 20°C for up to 14 days. For the fungus to grow in the tablets, they must have been exposed to an elevated temperature and humidity. Rhizopus microsporus is a ubiquitous fungus with the ability to grow at elevated temperatures up to 42°C and they are often the first fungal invaders of solid substrates. It forms bundles of rosette-forming sporangiophores bearing sporangia and individual ascospores. The most clinically important zygomycetes are Rhizopus species in the Order Mucorales⁷. Of 190 US clinical isolates tested, the frequency of species were Rhizopus oryae (45 per cent), R. microsporus (22 per cent), Mucor corymbifer (five per cent), Rhizomucor pusillua (four per cent), Cunninghamella bertolletiae (three per cent), M. indicus (three per cent), and C. echinulata (one per cent). The most common sites of infection were the sinuses (26 per cent), lungs (27 per cent) and various cutaneous locations (28 per cent).

Also Rhizopus microsporus is used in the Asian fermented food Tempe which is derived from soybeans. The optimum conditions for radial growth and biomass dry weight on defined media were temperature 40°C, water activity 0.995 and ambient air. At A_w <0.96, virtually no growth occurred⁸. This is inconsistent with R. microsporus growth in the granulation. A review of the literature strongly suggests R. microsporus growth in the Allopurinol compressed tablets occurred after manufacture when the compressed tablets were exposed for an extended period to elevated temperature and humidity. As the strain matching did not conclusively confirm the infection was derived from the tablets, it is possible that other environmental factors, e.g. unauthorised food like Tempe, were the source of the infections.

Skin moisturising lotion

A fungal-contaminated skin lotion was the probable cause of an outbreak of invasive mycoses in a haematology-oncology isolation and bone marrow transplantation unit of the University Hospital, Basel, Switzerland⁹. Twelve of 25 patients (48 per cent) admitted between 17 August and 31 October 1993 were infected with the fungus Paecilomyces lilacinus with skin eruptions while all five patients that were recipients of bone marrow grafts and four of 12 patients treated with chemotherapy for leukaemia and lymphoma developed invasive infections. After an extensive investigation of the unit environment, air handing systems, parenteral preparations, infusate bottles, transfusion sets and food, all topical agents administered to the patients were cultured. Skin lotion used in the general medical ward and the surgery department was found to contaminate the hands of a nurse and an ulcer of a patient on which the lotion used was found to be contaminated with the fungus P. lilacinus. Subsequently, 12 of 16 sealed bottles of the skin lotion were found to be contaminated with 6-12,500 cfu per mL of lotion. The manufacturer investigated with confirmation of intrinsic fungal contamination. The lotion consisted 36 per cent lipids, 40 mg/mL urea, and 0.3 per cent trilosan and 0.34 per cent chlorhexidine dihydrochloride as preservatives. The ingredients were not contaminated but P. lilacinus was found in the empty containers awaiting filling.

Wooden tongue depressors

Four cases of cutaneous infection with Rhizopus microsporus in vulnerable preterm infants in one neonatal nursery were reported by the Neonatal Unit, Birmingham Women’s Hospital, Edgbaston, UK¹⁰. The source of infection was identified as wooden tongue depressors, which were used on the nursery to construct splints for intravenous and arterial cannulation sites. The outbreak was ended by the removal of these items from the nursery. The report concluded that the combination of warm, humid conditions in neonatal incubators, particularly in association with occlusive dressings, may favour cutaneous fungal invasion and place premature babies at risk of infection.

In 1995 – 1996, a nosocomial outbreak of gastric mucormycosis caused by R. microsporus var. rhizopodiformis occurred in five adult patients admitted to a hospital intensive care unit in Pamplona, Spain¹¹. Wooden tongue depressors contaminated by R. microsporus var. rhizopodiformis used to prepare oral medications given to patients via nasogastric catheter caused this outbreak of fungal gastritis with an attributable mortality of 40 per cent. The authors concluded that wooden material should not be used in the hospital setting.

Sterile compounded methylprednisolone injections

On 21 September 2012, CDC was notified by the Tennessee Department of Health of a patient with the onset of meningitis approximately 19 days following epidural steroid injection at a Tennessee ambulatory surgery centre. Initial cultures of cerebrospinal fluid (CSF) and blood were negative; subsequently, Aspergillus fumigatus was isolated from CSF by fungal culture. On 28 September, investigators identified a case outside of Tennessee, possibly indicating contamination of a widely distributed medication. As of 4 October 2012, 35 people in six states – Tennessee, Virginia, Maryland, Florida, North Carolina and Indiana – had contracted fungal meningitis and five of them have died, according to the CDC. All received steroid injections for back pain, a highly common treatment. In response to the outbreak, the compounding pharmacy involved, the New England Compounding Center of Framingham, Massachusetts, recalled three lots consisting of a total of 17,676 single-dose vials of the steroid, preservative-free methylprednisolone acetate¹².

The Infectious Diseases Society of America recommends the antifungal agent Voriconazole treatment for Aspergillus osteomyelitis for a minimum of six to eight weeks for non-immuno-compromised patients with, as necessary, surgical debridgement and stabilisation of the spine. With these patients the death rate is 20 – 30 per cent, while it is around 100 per cent for immune-compromised patients. The contaminating fungus was subsequently redefined as Exserohilium rostratum, a fungus so rare that the Tennessee state health commissioner Dr John Dreyzehner described it as a fungus most physicians never see it in a lifetime of practicing medicine. According to the ASM Manual of Clinical Microbiology, phaeohyphomycosis of the skin, subcutaneous tissue, cornea, nasal sinuses and brain have been documented.

According to a 7 October 2012 article in the New York Times, starting in the 1990s, spinal injections for back pain, known as lumbar epidural steroid injections, skyrocketed. They have since levelled off, but the number remains high. In 2011, 2.5 million Medicare recipients had the injections, as did an equal number of younger people, according to Dr. Ray Baker, president of the International Spine Intervention Society.

Sterile compounding pharmacies are regulated by their State Pharmacy Boards and not the FDA. Their legal requirements are found in USP <797> Pharmaceutical Compounding – Sterile Preparations. Instead of compounding a drug product in response to a physician’s prescription, large-scale compounders often behave like manufacturers, complete with sales teams that market their products to doctors. As compounding pharmacies, they do not have to abide by the FDA’s GMP regulations, which require that problems with products be reported to the agency circumventing the regulatory process. In October 2012, the State of Massachusetts and the federal justice department started a criminal investigation into the New England Compounding Center.

On 26 October 2012, the FDA completed an inspection of the New England Compounding Center and issued an eight page Form 483 inspectional observations report. The inspection was against the GMP regulations as applied to sterile product manufacturing facilities. According to an industry summary the inspection (www.examiner.com/article/necc-had-issues-with-clean-room-contamination-fda), it revealed issues with bacterial and fungal contamination in the clean rooms. The environmental monitoring records show the cleanrooms and ancillary rooms and areas had counts of bacteria and moulds that frequently exceeded the action level.

Dozens of samples of the drug implicated in the multistate fungal meningitis outbreak, the injectable steroid, methylprednisolone acetate contained either greenish black foreign matter or white filamentous material. Sterility testing confirmed the presence of fungi. Discolourations and condensation were observed on several pieces of equipment used at the facility like autoclaves used in the manufacture of sterile product, including methylprednisolone acetate. In addition, dark particulates and white filamentous substances covered the louvers of the Heating Ventilation and Air Conditioning (HVAC) return behind the autoclaves. Inspectors also noted that large equipment used for excavation in a waste recovery area was producing airborne particulates outside the facility, approximately 100 feet from NECC’s HVAC system.

To assist with the rapid detection of the fungal DNA including that of Exserohilium rostratum in clinical specimens, CDC scientists developed a PCR test that uses the amplification and sequencing of ribosomal ITS2 using the ITS3/4 broad range primer sets and Exserohilium-specific primers¹³.

As of 28 January 2013, the contaminated Methylprednisolone acetate distributed by NECC has resulted in 693 cases of fungal infection and 45 deaths with nearly 14,000 potentially-exposed patients across 23 states¹⁴.

Water activity measurement (non-sterile products)

The water activity of a non-sterile pharmaceutical drug product is the most critical physical attribute that determines whether a specific drug product will support the growth of microorganisms. USP Informational chapter <1112> Application of water activity determination to non-sterile pharmaceutical products references the dew point / chilled mirror method contained in AOAC International Official Methods as a suitable method for water activity determination.

Water activity ranges from 1.00 for pure water to 0.00 for a bone dry material. As most microorganisms require a water activity of at least 0.75 for microbial growth, the proliferation of microorganisms within a drug product can be controlled by reducing the water activity of the drug product. In general, fungi out compete bacteria as the water activity is lowered. For example, syrups with a water activity of 0.95, if unpreserved, would support microbial growth while a topical ointment with a water activity of 0.58 would not support growth (Table 4).

Water activity determination should be a routine part of formulation development and may be used as tool for microbiological risk assessment and specification setting¹⁵.

Table 4: The Water Activity and Limits for Microbial Growth associated with Different Dosage Forms

Methods used for the identification of fungi

Most mycological laboratories have relied on phenotypic identification using colony morphology, colour and sporulation, cellular diagnostic features and carbohydrate utilisation patterns that are time-consuming and subjective and depend largely on the skill and experience of the mycologist in that laboratory¹⁶. The application of rapid rDNA base sequencing methods results in more timely and accurate species-level identification that improves clinical outcomes. There are two major sequencing targets for fungal identification. They are the D1/D2 region of the large ribosome subunit (LSU) and the internal transcribed spacer regions (ITS1/ITS2). Of late the ITS sequence has become the preferred target for fungal identification¹⁷. Mycologist estimate there are of the order of 1.5 million fungal species with less than one per cent of this number sequenced for the ITS region. Even within the 165,000 fungal ITS sequences in the International Nucleotide Sequence Databases (ISND: GenBank/European Molecular Biology Laboratory (EMBL)/DNA Database of Japan (DDBJ) less than half have full species names and an estimated 10 per cent have incorrect names¹⁸. Simply put, to identify a fungal isolate it’s amplified ITS sequence is compared to valid ISND sequences using BLAST for the relevant matches.

For sequencing analysis, the genomic DNA is extracted from the culture, the target gene fragment is PCR amplified using specific PCR primers and thermocycling conditions, the PCR amphlicon is purified and the order of the nucleotide bases in the DNA target is determined by base sequencing¹⁹. The dye-terminator sequencing method, using automated DNA sequence analysers (Applied Biosystems, Foster City, CA), is used for most base sequencing. As discussed by Balajee et al.²⁰, the success a sequencing strategy for fungal identification depends on the following:

• Choice of the fungal DNA region subjected to PCR amplification, i.e. D2 or ITS regions
• amenability of that region to amplification and sequencing for a wide range of fungi
• reliability of the interpretation of the results
• availability of a suitable proprietary or public sequence database for comparison.

Evaluations of the MicroSeq D2 Large-Subunit rRNA sequencing^21,22 and the ITS D1/D2 rRNA sequencing for fungal identification^23-25 have been published in the clinical microbiology literature. The most commonly used fungal identification methods and the relative sizes of their databases are summarised in Table 5.

Given the diminished mycological expertise within our industry, slow fungal growth and lack of diagnostic morphological features on some fungal isolates, we can expect that microbiologists will more widely employ phenotypic and/or proteomic techniques supplemented by rRNA base sequencing to identify fungi.

Table 5: Comparison of Methods and Databases of Common Phenotypic and Genotypic Fungal Identification Systems

Conclusions

As stated in the introduction, the author believes that more attention should be given to the mitigation of the risk of fungal contamination of drug products and medical devices. A chronic but low level of product recall for fungal contamination highlights this risk. Manufacturers should better control fungi within their facilities, formulate products with more robust antimicrobial preservative systems and establish risk-based product testing programs to detect fungi in their products. As microbial identification is a key element in failure investigation companies should upgrade their methods to include proteomic methods and rRNA base sequencing.

Investigations of microbial infection outbreaks associated with contaminated drug products and medical devices have revealed a pattern of infection of immune-compromised patients intensive care facilities from diverse sources including wooden tongue depressors, alcohol wipes, probiotics, disinfectants, lubricating gels, and skin lotions. It appears that many of these materials are not formally charted on patient records and are only uncovered by astute and comprehensive internal investigations. Greater awareness is needed from hospital staff in terms of the potential impact of the use of contaminated non-sterile drug and consumer health products and medical devices in intensive care units.

References

1. Perlroth, J., B. Cho and B. Spellberg, 2007 Nosocomial fungal infections: epidemiology, diagnosis and treatment. Med. Mycol. 45:321-346
2. Richardson, M. D. and D. W. Warnock, 2012 Fungal Infection: Diagnosis and management Fourth Edition Wiley-Blackwell pp445
3. Repetto, E.C., C.G. Giacomazzi and F. Castelli 2012 Hospital-related outbreaks due to rare fungal pathogens: a review of the literature from 1990 to June 2011 Eur. J. Clin. Microbiol. Dis. 31:2897-2904
4. Sutton, S and Jimenez, L 2012. A Review of reported results involving microbiological control 2004-2011 with emphasis on FDA consideration of “objectionable organisms.” American Pharmaceutical Review 15(1), January 2012.
5. Chin-Hong, P 2006 Infections in patients with diabetes mellitus: Importance of early Recognition, treatment and prevention Adv Stud Med. 6(2):71-81
6. Cheng, V.C.C., J. F. W. Chan, A. H. Y. Ngan, K. K. W. To, S. Y. Leung, H. W. Tsoi, W. C. Yam, J. W. M. Tai, S. S. Y. Wong, H. Tse, I. W. S. Li, S. K. P. Lau, P. C. Y. Woo, A. Y. H. Leung, A. K. W. Lie, R. H. S. Liang, T. L. Que, P. L. Ho, and K. Y. Yuen, 2009 Outbreak of intestinal infection due to Rhizopus microsporus J. Clin. Microbiol. 47 (9) 2834-2843
7. Alvarez, E., D. A. Sutton, J. Cano, A. W. Fothergill, A. Stchigel, M. G. Rinaldi, and J. Guarro1, 2009. Spectrum of zygomycete species identified in clinically significant specimens in the United States. J. Clin. Microbiol. 47 96) 1650-1656
8. Han, B.Z. and M. J. R. Nout, 2000. Effects of temperature, water activity and gas atmosphere on mycelial growth of tempe fungi Rhizopus microsporus var. microsporus and Rhizopus microsporus var. oligosporus. World J Microbiol. & Biotechnol. 16:853-858
9. Orth B., R. Frei, P.H. Itin, M.G. Rinaldi, B. Speck, A. Gratwohl and A.F. Widmer, 1996 Outbreak of invasive mycoses caused by Paecilomyces lilacinus from a contaminated skin lotion Ann. Intern. Med. 125 (10): 799-806
10. Mitchell S.J., J. Gray, M.E.I. Morgan 1996 Nosocomial infection with Rhizopus microsporus in preterm infants: association with wooden tongue depressors lancet 348: 441-443
11. Maravi-Poma, E., J.L. Rodriguez-Tudela and J.G. de Jalon 2004 Outbreak of gastric mucormycosis associated with the use of wooden tongue depressors in critically ill patients Intensive Care Med. 30:724-728
12. Kauffman, C. A., P. G. Pappas and T. F. Thomas, 2013 Fungal Infections associated with contaminated Methylprednisolone injections NEJM 368: 2495-2500
13. Gade, L., C. M. Scheel, C. D. Pham, M.D. Lindsley, N. Iqbal, A. A. Cleveland, A. M. Whitney, S. R. Lockhart, M. E. A. P. Litvintseva, 2013 Detection of fungal DNA in human body fluids and tissues during a multistate outbreak of fungal meningitis and other infections. Eukaryotic Cell 12 950: 677-683
14. Bell, B.P. and R. F. Khabbaz, 2013 Responding to the outbreak of invasive fungal infections: The value of public health to Americans. JAMA 309 (6): 883-884

1. Cundell, A. M., 2009 Chapter 9: Effects of Water Activity on Microorganisms In Cundell, A. M and A Fontana (Editors) Water Activity Application in the Pharmaceutical Industry Davis Horwood/PDA 2009

16. Larone, D.H. 2002 Medically-important Fungi: A Guide to Identification. Fourth Edition ASM Press, Washington, D.C.
17. Seifert, K. A., 2009 Progress towards DNA barcoding of fungi. Mol. Ecol. Res. 9: 83-89
18. Nelsson, R. H., K. Abarenkov, K-H. Larsson and U. Koljalg, 2011 Molecular identification of fungi: rationale, philosophical concerns and the UNITE database. Open Appl. Inform. J 5 (Suppl. 1-M9) 81-86
19. Dong J., M.J. Loeffelholz and M. R. McGinnis 2012. Sequence-based fungal identification and classification. In Molecular Microbiology; Diagnostic Principles and Practice Second Edition ASM Press pp669-676
20. Balajee, A.R., L. Sigler and M. E. Brandt, 2007 DNA and the classical way: identification of medically important molds in the 21^st century. Med. Mycol. 45:475-490
21. Rozynek, P. S. Gilges, T. Bruning and M. Wilhelm, 2003 Quality test of the Microseq D2 LSU fungal sequencing kit for the identification of fungi. Int. J. Environ. Health 206: 297-299
22. Hall, L., S. Wohlfiel and G. D. Roberts, 2004 Experience with the MicroSeq D2 large-subunit ribosomal DNA sequencing kit for identification of filamentous fungi encountered in the clinical laboratory J. Clin. Microbiol. 42 (2): 622-626
23. De Baere, T., G. Claeys, D. Swinne, C. Massonet, G. Verschragen, A. Muylaert and M. Vaneechouttel, 2002 Identification of cultured isolates of clinically important yeast species using fluorescent fragment length analysis of the amplified internally transcribed rRNA spacer 2 region. BMC Microbiol. 2:21
24. Landlinger, C., L.Baskova, S. Preuner, B. Willinger, V. Buchta and T. Lion , 2009 Identification of fungal species by fragment length analysis of the internally transcribed spacer 2 region. Eur. J. Clin. Microbiol. Infect. Dis. 28: 613-622
25. Kwiatkowski, N. P., W. M. Babiker W. G. Merz, K. C. Carroll and S. X. Zhang, 2012 Evaluation of nucleic acid sequencing of the D1/D2 region of the large subunit of the 28S rDNA and the Internal transcribed Spacer region using SmartGene IDNS software for identification of filamentous fungi in a clinical laboratory. J. Mol. Diagnos. 14 94): 393-401