Diagnosis and Management of Pneumocystis Pneumonia in Resource-poor Settings
Globally, Pneumocystis pneumonia (PCP) remains a common and lethal infection in both HIV-positive and HIV-negative patients, particularly in developing countries where rates of PCP increases with rising GDP. Pneumocystis jirovecii cannot be cultured in routine clinical laboratories; thus diagnosis relies on microscopy, histology, serology and/or polymerase chain reaction (PCR) of the Pneumocystis DNA. Most of these methods are expensive and require training. Accessing lower respiratory tract specimens in young children is often challenging and only PCR testing of nasopharyngeal aspirates is useful. Early treatment with high-dose co-trimoxazole is effective therapy; however, adverse reactions are common along with reports of emerging resistance. Improved outcomes are associated with adding corticosteroid to treatment in those with moderate/severe PCP, although this has not been studied in resource-poor settings. This review compares the available diagnostic techniques in relation to their suitability for use in resource-poor settings. We also addressed the non-availability of the alternative medications in these regions.
Pneumonia, Pneumocystis, HIV, diagnostic tests, children, Africa, low-middle income countries.
Among people with AIDS worldwide, Pneumocystis pneumonia (PCP) remains a common and life-threatening opportunistic infection.1 With the increased use of chemotherapeutic agents and immunosuppressants, the incidence of PCP among patients without HIV infection has progressively increased, and is associated with mortality rates of 35–55% compared with 10–20% among HIV-infected patients.2 A rough estimate of the annual incidence of PCP among people with AIDS is more than 400,000 adults and [End Page 107] children worldwide (~14.7% of 1.76 million with CD4 cell count <100).3 Pneumocystis pneumonia has increased as gross domestic product (GDP) increases globally.3
Pneumocystis jirovecii (formerly carinii) is a ubiquitous organism and it is estimated to infect as high as 95% of the worldwide population during the first two years of life with asymptomatic carriage in healthy people.4 P. jirovecii is specific to humans with multiple genotypes.5 Early diagnosis and treatment with high dose co-trimoxazole is effective therapy; however, adverse reactions are very common, notably: nausea (>90%) and vomiting, rash 35%, neutropenia (55%), 25% fall in haemoglobin (20%), and abnormal liver function tests (2–5x rise in enzymes) (35%).6 A meta-analysis revealed that improved outcomes are associated with adding corticosteroid to treatment regimens in those with moderate or severe PCP.7 Resistance is rare but has been reported and is associated with dihydropteroate synthase polymorphisms.8,9 In resource-poor settings where empirical therapy is advocated (due to lack of diagnostics) many patients who do not have PCP are unnecessarily exposed to high dose co-trimoxazole. Pneumocystis pneumonia morbidity and mortality remains high in HIV-positive individuals who do not have access to antiretroviral therapy (ART) and in whom ART tolerance is an issue or the drugs are ineffective. It is also high in HIV-positive people who are ignorant of their HIV status and those who, due to fear of stigmatisation, choose not to seek medical care.10
A multicenter, 8-year prospective study to determine the impact of HIV-associated conditions on mortality showed an overall mortality rate of 5.41 deaths per 100 person-years, and PCP was associated with a doubling of mortality.11 Presently, P. jirovecii cannot be cultured in routine clinical laboratories so diagnosis rests on microscopy of respiratory fluids, histology of lung tissue, serological biomarkers and/or polymerase chain reaction (PCR) detection of Pneumocystis DNA. (See Figure 1.) More sensitive than microscopy (~98% versus 75% for silver stains), PCR has the advantage in resource-poor settings of requiring minimal training; however it is expensive.12 Only moderate to severe cases are readily diagnosed based on clinical and radiological findings, thereby missing mild cases. A high performing diagnostic assay would allow earlier treatment of mild cases and discontinuation of therapy in those without PCP. The cost-effectiveness of diagnostic assays improves in places where there is a high disease prevalence.
This review focuses on the epidemiology, challenges with laboratory diagnosis and therapeutic management of PCP in relation to resource-poor settings.
Search strategy and selection criteria
The literature search for publications on diagnosis and management of PCP preceding 30 March 2016, was performed using PubMed (accessed MEDLINE), Web of Science, Google Scholar, Cochrane Library, African Journals Online (AJOL), Africa-Wide: NiPAD, CINAHL (accessed via EBSCO Host) databases and grey literature to identify all published papers regarding the topic. The references were reviewed for additional publications that may not have been cited elsewhere ("snowballing"). Articles published in other languages (e.g., French and Portuguese) were considered if they were cited in any of the databases searched. The main search comprised individual searches using detailed medical subject heading (MeSH) terms for pneumocystis pneumonia, community-acquired pneumonia and HIV/AIDS combined with terms relevant to PCP diagnosis and management. The Boolean operators 'AND' and 'OR' were used to combine and narrow the searches. [End Page 108]
In 1981, two case reports of PCP in five previously healthy homosexual males who were injection drug users announced the beginning of the HIV/AIDS pandemic.13,14 Pneumocystis pneumonia is the commonest AIDS defining opportunistic infections in HIV infected people in the U.S. and Europe,15,16 but it was initially assumed to be rare in low and middle-income countries (LMIC), such as African countries.17,18 However, more recent studies have reported contrary findings.19–21 There are plausible reasons for the low PCP rates previously reported in LMICs. One is the widespread poverty combined with low quality of health care that may result in most HIV infected patients dying from infection before they can develop PCP; another is the lack of diagnostic facilities and trained personnel to identify Pneumocystis in most of these countries.
Studies from Asia reveal varying rates of PCP in the last two decades, ranging from18.7–25.4% in HIV infected patients in Thailand, with attendant high mortality,22–26 and 16.7% in Bangladesh,27 8.4% in Cambodia28 and 5% in Vietnam.29 Earlier data from India demonstrated rates of 5–6.1% of PCP in HIV-infected individuals.30–32 However, with better molecular detection techniques, higher PCP rates of 12.2–26.5% are being reported in India.33–35 A recent study from India reported an incidence of 14% in 94 immunocompromised children of whom 14 were HIV-infected.36
In South America, the picture is practically the same, although there is paucity of data. Studies on PCP in HIV-infected people there revealed a 24% incidence rate in Mexico,37 48% in Panama,38 27% in Guatemala,39 32% in Cubans,40 and 35% in Haiti.41 [End Page 109] Other South American reports demonstrated 36.6% from Venezuela,42 Peru 12.5%,43 38% in Chile,44 and 27% in Brazil.45
Burden in Africa
In Africa, PCP was previously assumed to be uncommon among the HIV population.18,46–49 Early studies from Uganda and Zambia reported no cases of PCP among HIV-infected patients.17,49 A South African report revealed similar findings of one (0.6%) positive sample out of 181 patients tested for PCP.50 However, in the same period an incidence of 3.6–11% was documented among HIV-infected people in Tanzania, Congo and Ivory Coast.18,51–53 Additionally, in a setting that had better diagnostic facilities and increased access to ART, a PCP prevalence of 33% from 64 smear negative tuberculosis (TB) patients in Zimbabwe using methenamine silver staining on bronchoalveolar lavage (BAL) samples was reported.54 Another report from Kenya, using immunofluorescence (IF) and toluidine blue staining identified Pneumocystis in 37.2% and 27.4% respectively of 51 HIV/AIDS infected patients.55 In an Ethiopian report, P. jivovecii was detected by PCR in 42.7% of 131 BAL samples from HIV-infected patients with atypical radiological reports who were acid fast bacilli (AFB) smear negative56 and 29.7% by IF.57 In Nigeria, 12.6% was reported positive using Pneumocystis PCR.58
Studies in children showed similar results and PCP appears to occur early among HIV-infected infants (median age: approximately 13 months), suggesting that exposure to Pneumocystis is relatively extensive. A recent study from Mozambique demonstrated a 6.8% prevalence of PCP with 14.3% in HIV-infected children and 3.3% in non HIV-infected children.59 At the start of the HIV/AIDS pandemic, the incidence of PCP was 1.3 cases per 100 child-years from early childhood to adolescence and went up to 9.5 cases per 100 child-years in infancy.60,61 Postmortem studies of lung tissues from children with AIDS revealed an incidence of 67% in Zimbabwe;21 31% in children younger than 15 months old in Ivory Coast62 and 48% in HIV infected children under 12 months in Botswana.63 One of the challenges with diagnosis in infancy is that age three-six months has been shown to be a period of high incidence of PCP.4 However, the child's HIV status is usually undetermined at that period in most resource-poor settings because these patients are not routinely presented for care.64 Anti-Pneumocystis antibodies were demonstrated in HIV-negative children in early years of life (aged 1.9–19 months; mean, 7.1 months; median, 5 months; SD, 4.9)65 and as early as two-six months in African children, often with it being the first presentation of HIV related disease.66 Following improvement of prenatal HIV testing and introduction of ART to prevent vertical spread, there has been a significant decrease in paediatric HIV infections. The incidence of PCP also reduced substantially in children from 1992 to 1997, with a sharp decline from 1995 and this was attributed to improving ART administration in labour.67 Despite this, a study from Mozambique among children younger than five years of age reported a prevalence of 6.8% in newly presenting children with severe pneumonia, of whom 25.7% had HIV infection and 59% of the PCP cases were in those with HIV infection.59 Table 1 shows the distribution across resource limited countries.
Outbreaks of PCP suggestive of human transmission were first documented among hospitalized oncology and transplant patients in United States and Europe.68–71 These were followed by reports of outbreaks among AIDS patients and [End Page 110]
[End Page 111] immunosuppressed rheumatoid arthritis patients.72–75 The possibility of transmission of P. jirovecii to and by health care workers (HCWs) has also been investigated with some studies reporting substantial differences in antibody titer levels in HCWs exposed to PCP.76 Another study demonstrated a significant increase in those that been exposed to PCP, keeping in mind that this is an aerosol transmitted disease.76,77 A study measuring levels of antibodies to the major surface glycoprotein (Msg) of Pneumocystis demonstrated higher levels in HCWs exposed to PCP than in non-HCWs that were not exposed to the infection78 implying that HCWs can serve as a reservoir for P. jirovecii. These reports pose a challenge in the management of PCP patients considering that current international guidelines do not advocate respiratory isolation for these patients. Single room isolation for PCP to minimize transmission is desirable for the first week of therapy but not realistic in most LMICs. [End Page 112]
Pneumocystis, the organism
Pneumocystis is an ubiquitous, obligate, biotrophic, extracellular eukaryotic organism that exists in trophic and cystic forms.79 It has unique mechanisms of adaptation to life exclusively in mammalian species and is known to be host-species specific.80 Ma and colleagues in 2016 reported successful genome analysis of three Pneumocystis species (human, rat and mice) which demonstrated that adaptation mechanisms occurs exclusively in mammalian hosts lungs.81 Humans act as a reservoir for P. jirovecii; however, the precise association is not fully understood and environmental reservoirs have also been documented outside the lungs.82–84 Pneumocystis was first grouped with protozoans, but in 1988 it was reclassified as a fungus because its ribosomal RNA was most similar to that found in fungi.79,85,86
Pneumocystis spp. does not appear to grow in standard fungal culture, although it can be detected in the environment by molecular methods.87 However, a recent study from Germany described an innovative method to culture P. jirovecii using differentiated pseudostratified CuFi-8 cells that were inoculated with BAL fluid (confirmed positive by PCR for P. jirovecii).88 Although the efficacy of such a culture system for propagating the organism and/or directed therapy selection is yet to be determined, it is nevertheless an innovation that will affect the diagnosis and management of PCP.
Transformation from cyst to trophozoite
Pneumocystis appears to have a bi-phasic life cycle within the alveolar lumen, consisting of an asexual phase characterized by binary fission of trophic forms and a sexual cycle resulting in formation of cysts.89 The trophic form possesses a single nucleus and has a flexible cell wall that is often tightly attached to type I pneumocytes in lung alveoli. The cyst can contain up to eight 'sporozoites' and is protected by a distinctive thick cyst wall. When the wall ruptures (encystment), these 'sporozoites' are released and then develop into new trophic forms.90–92 The cell wall of the organism, both cystic and trophic forms, contains melanin-like compounds which protects it from environmental stressors.93 The cyst wall is maintained by a unique system of building and breaking down which sustains its rigidity and viability, may reduce immune recognition and also ensures the organism completes its life cycle. β-glucan synthetases are the enzymes involved in production of β-1,3-glucan homopolymers that make up the cyst wall; β-glucanases and other enzymes drive the active process of encystment.94–98 These enzymes are noteworthy as targets for the development of new therapeutic molecules.
HIV and other risk factors
A number of factors are associated with the development of PCP but impaired T-cell immunity is the pivotal risk factor for PCP.10,92,99 Sustained defective immunity from past immunosuppressive therapy, a number of immunosuppressive conditions such as haematological malignancies especially leukemia and lymphomas, solid organ cancers and transplants are known risk factors, as well as a wide variety of specific immunosuppressive medications.100 Other patient groups at risk include post-transplant patients, those with autoimmune and inflammatory conditions such as rheumatologic and other anti-inflammatory processes, particularly when they are exposed to prolonged high dose corticosteroid therapy.92,101 Cytotoxic drugs such as methotrexate, cyclosporine and cyclophosphamide have been implicated in the development of PCP.102–104 Additionally, newer immunomodulating agents such as TNF-a inhibitors have also been associated with this disease.105,106
The most substantial risk factor for PCP in HIV infected patients is a CD4+ cell count [End Page 113] below 200 cells/ml.10,99,107–109 CD4+ cells are necessary for the clearing of Pneumocystis. This has been revealed both in experimental models, where a direct relationship between CD4+ cell count below 200 cells/ml and the development of PCP infection was shown.108 Fortunately, the advent of ART, which boosts recovery of CD4+ cell count levels, has led to significant reduction in rates of PCP in HIV-infected patients.87 This seems to be the case in industrialized countries where the majority of HIV positive patients have access to ART. However, the contrary is the case in resource-poor countries.
A recent systematic review showed that the most significant predictor of PCP was per capita GDP, which demonstrated strong linear association with odds of PCP diagnosis (p <.0001).3 It is likely that poverty exposes HIV-infected patients to a variety of pathogenic organisms with P. jirovecii being just one of many in LMICs. However, as the economics of the LMICs improve, it is plausible that many virulent bacterial infections circulate less frequently and in HIV-infected people PCP may assume a greater role. The surrounding indoor air is also important in PCP transmission. A study of hospitalized PCP patients in France demonstrated P. jirovecii DNA in 80% of air samples in the patients' immediate surroundings, with progressive reduction with increasing distance from patient's bedside.110 Other researchers have reported the detection of DNA sequences identical to P. carinii in samples of ambient air.111–113 Yet another study highlighted P. jirovecii exhalation from colonized patients and emphasized the risk of nosocomial transmission of the disease.114 Morris and colleagues113 found that geographical location is another factor associated with PCP and other researchers have confirmed this finding.19,115–117 Several studies have also assessed the role of seasonal variation in PCP with conflicting results, with some peaks of the infection in summer or winter seasons or no seasonal variation at all.116–118 It will be interesting to see if studies from tropical regions such as Africa will reflect seasonal variations.
Impact of prophylaxis on incidence
The Centers for Disease Control and Prevention (CDC) HIV/AIDS surveillance reports of 1990–1993 in pre-ART and prophylaxis era showed that PCP accounted for over 20,000 new AIDS cases yearly in the United States.119 With the institution of ART and PCP prophylaxis, the rates of infection dropped from 31% to 9% in the U.S.120 In Europe, PCP accounted for 16.4% of AIDS cases,121 though most of the cases were in ART naïve patients. In a EuroSIDA study, 7333 HIV-infected people were recruited from 52 centres across Europe and Israel, and the PCP rates dropped from 4.9 cases per 100 person-years to 0.3 cases per 100 person-years after the use of ART was consolidated.122 Most recent data from these countries suggest that PCP appears mainly among HIV-infected patients that do not know their HIV serostatus, or who have challenges with ART or PCP prophylaxis compliance.67,123–125
In the HIV-infected population, the diagnosis of PCP presents a clinical dilemma because there are no specific signs and symptoms of the disease. There is also the challenge of co-existing morbidities with other pathogens and the use of prophylactic drugs in managing these patients.12,126,127 These patients usually present with sub-acute onset of gradual dyspnoea, nonproductive or minimally productive cough, low grade fever and malaise. Early in the course of infection patients may be asymptomatic. Acute dyspnoea with associated pleuritic chest pain is likely indicative of pneumothorax complicating PCP. In contrast, non-HIV immunosuppressed patients tend to present more acutely, with significant dyspnoea, high fever, chills and [End Page 114] in some cases with respiratory failure which could result in up to 40% mortality.129 In children, in addition to dyspnoea and fever, they may have cyanosis, nasal flaring, and intercostal retractions. Physical examination in both adults and children tends to reveal tachycardia, tachypnoea and a 'clear chest' on auscultation, but sometimes inspiratory crackles are heard.99 Pneumocystis pneumonia is graded based on clinical features into mild, moderate and severe to aid in management of patients (see Table 2; see Table 3 for comparison of diagnostic methods).
HIV-infected PCP patients generally have more Pneumocystis organisms with fewer neutrophils in their bronchoalveolar lavage (BAL) specimen than non-HIV immunosuppressed patients.99 This greater burden of infecting organism correlates with a significantly higher diagnostic yield.99 However, the smaller number of inflammatory cells does not seem to result in worsening oxygenation or impact on survival. In fact, the opposite tends to occur. HIV infected PCP patients also appear to have higher arterial oxygen tension and a lower alveolar-arterial oxygen gradient than non-HIV immunosuppressed PCP patients.129 This probably explains why non-HIV immunosup-pressed PCP patients are more likely to develop respiratory failure than HIV infected PCP patients.129
Chest radiographic findings in PCP are nonspecific, and as many as one third of infected patients may have normal radiographic findings.130 High-resolution computed tomography (HRCT), which is more sensitive than chest radiography, is used when the chest radiograph is appears normal but there is a high index of suspicion or when the diagnosis is unclear. The HRCT may reveal extensive 'ground-glass' appearance or cystic lesions reflecting accumulation of intra-alveolar fibrin, debris, and organisms in PCP.130 High-resolution computed tomography is expensive and not cost-effective in LMICs.
Granulomatous reactions (see Figure 2) are estimated to occur in 5% of HIV-infected patients with PCP and usually occur early in the course of the infection when immunodeficiency is more reduced; they may become evident on HRCT as a single [End Page 115]
nodule or mass mimicking lung carcinoma or as multiple nodules ranging from a few millimeters to more than 1 cm.131 Cystic lesions may also be seen in PCP in AIDS; Tokuda and colleagues found that the presence of AIDS was found to be a risk factor for the formation of pulmonary cystic lesions using multivariate analysis.132 Cysts are associated with an increased frequency of spontaneous pneumothorax, however, spontaneous pneumothorax can also occur in the absence of definable lung cysts.133 These cysts may resolve after drug therapy and resolution of infection.131 [End Page 116]
Other radiological features include diffuse bilateral interstitial infiltrates, solitary or multiple nodules, upper-lobe infiltrates in patients receiving aerosolized pentamidine, pneumatoceles, pneumothorax and patchy asymmetric infiltrates.64,134–136 In more advanced disease, septal lines with or without intralobular lines superimposed on ground-glass findings ('crazy paving')137 and consolidation may develop.130 A recent case report showed symmetric bi-apical cystic spaces in chest radiographs.138 These air spaces may subsequently get infected with Aspergillus spp and an aspergilloma may form.139–142 Patients recovering from PCP may have residual interstitial fibrosis.143 In addition, although rare, interstitial fibrosis can occur in AIDS patients with low-grade chronic PCP, a condition termed chronic Pneumocystis pneumonia.144
Extrapulmonary PCP does occur rarely in patients with advanced AIDS.145–150. Some of these manifestations include: pneumocystic lesions of bone, brain, kidney, liver, spleen, eye, thyroid and the gastrointestinal tract.145–147,149,151–153
While severe and typical PCP can be diagnosed clinically, survival is better with earlier diagnosis.154 Perhaps the greatest benefit of laboratory testing for PCP is to rule out the diagnosis of PCP thereby avoiding unnecessary exposure to toxic doses of cotrimoxazole. Early diagnosis may also prevent hospital admission in mild/moderate cases.
A variety of specimens have been used for diagnosis, including lung biopsy, bronchoalveolar lavage (BAL), induced and expectorated sputum, nasopharyngeal aspirates and oral washings. The high morbidity associated with biopsy specimens has limited their clinical utility so BAL has largely been the sample of choice.155,156 Bronchoscopy is required for lung biopsy and BAL; its unpleasantness, cost, invasiveness and expertise required renders it impractical in resource-poor settings and so induced sputum has been preferred.157–161 Expectorated sputum is also a useful diagnostic specimen, and obviates the need for sputum induction.55,162,163 One study reported a 55% detection sensitivity for P. jirovecii in expectorated sputum,162 which was similar to the sensitivity with induced sputum.159,161 A retrospective review of PCP cases in the United States also found no significant difference in P. jirovecii yield between induced or expectorated sputum.164 Moreover, comparative evaluations of the general specimen quality of induced and expectorated sputum concluded that sputum induction did not improve the specimen quality substantially.165,166 In a recent study on expectorated sputum from HIV-infected and smear-negative TB patients in Namibia; of 475 samples analysed, 5.3% samples were positive for P. jirovecii, (13.6% using both qPCR and GMS staining and 1.7% using qPCR only).163 The study demonstrated that both standard microscopy with silver staining and real-time PCR were frequently positive and approximately concordant, indicating that expectorated sputum is probably a good specimen for PCP diagnosis.163
Physiotherapeutic interventions can safely and effectively procure a greater yield of sputum from patients that have difficulty producing sputum normally.167 However, sputum induction may be unsafe in some patients, particularly infants and weaker AIDS patients, because of the risk of haemoptysis in patients with tuberculosis or chronic pulmonary aspergillosis and the health care worker is potentially exposed to M. tuberculosis.167 It cannot be performed in children less than about 4 years old, because they don't understand or follow instructions, and usually swallow anything they [End Page 117] cough up. Therefore, correct detection of Pneumocystis presents a myriad of challenges especially in resource-poor settings. Though the procedures to obtain oral washes,168–170 nasopharyngeal aspirates171 and sputa160,172 are relatively less invasive compared with bronchoalveolar lavage, and are suitable for diagnosis with molecular methods but not so well with microscopy
Non-specific markers of PCP—lactate dehydrogenase (LDH)
Increased serum lactate dehydrogenase levels have been documented in patients with PCP in some studies but it is most probably due to the underlying lung inflammatory responses and damage rather than a precise biomarker for PCP.173,176 Lactate dehydrogenase has a high sensitivity for PCP but it is not specific.173
Direct microscopy of respiratory samples has been the gold standard for diagnosis of PCP. However it has limitations as inspection of stained slides is subjective and may be non-specific; sensitivity relies heavily on the organism load in the specimen, the type of sample collected and the expertise of the microscopist visualizing the slide. These challenges have led clinicians managing at risk patients to rely on radiological findings and clinical evaluations for diagnosis of PCP.
There are several useful methods for Pneumocystis detection on all sample types including immunofluorescence microscopy utilizing monoclonal antibodies; cyst wall stains (toluidine blue O, cresyl echt violet and calco-fluor white), and trophic form stains (modified Papanicolaou, GramWeigert, Grocott's methenamine silver stain, Diff-Quick, Wright or Giemsa). These different methods have their advantages and challenges (see Table 4). A comparison of three stains (IFA, Diff-Quik and Toludine blue O) demonstrated a sensitivity of 92% for IFA, 76% for Diff-Quik and 80% for toluidine blue O with no false positives for IFA.177 Another study by Procop and colleagues documented that "the sensitivity and specificity of calcofluor white stain (CW) were 73.8 and 99.6%, respectively; that of Grocott-Gomori methenamine silver stain (GMS) were 79.4 and 99.2%, respectively; Diff-Quik stain (DQ) were 49.2 and 99.6%, respectively while that of the Merifluor Pneumocystis stain were 90.8 and 81.9%, respectively.178 Only CW and GMS had positive and negative predictive values of >90%.178 Calcofluor white is quick, convenient, and can detect simultaneously the presence of other fungi in samples but expertise is required for identification of cysts of Pneumocystis so this stain may miss severely infected patients. All of these microscopy methods can have false-negative results, especially in samples from non-HIV immunosuppressed patients who have fewer organisms present in the specimens.129 Figure 3 demonstrates Pneumocystis using special stains.
Over 75 studies, most using in-house Pneumocystis assays, attest to the superiority of PCR amplification assays over microscopy both in sensitivity and in being able to analyse all sample types for the presence of P. jirovecii DNA. The advent of real-time quantitative PCR (qPCR) has enabled the rapid and contamination-free molecular diagnosis of PCP and several companies have commercialized tests. However, PCR is not technically and financially viable for resource-poor settings where the burden of this disease is high. A reliable electricity supply is often problematic, as is shipping reagents in dry ice through customs. A cost-effectiveness analysis by the Center for Disease Control and Prevention concluded that for the detection of [End Page 118]
[End Page 119] Pneumocystis, use of PCR assays, combined with less-invasive patient specimens such as nasopharygeal aspirate (NPA), oral washings, expectorated or induced sputum, represent more cost-effective alternatives than diagnostic techniques using BAL, or radiological findings alone.12
Polymerase chain reaction assays have also been used to analyze oral wash samples in the search for more non-invasive techniques.168,179,180 The generally higher sensitivity of these assays when compared to microscopy could be of greater value among non-AIDS patients in whom the fungal load is generally lower.181–184 These more sensitive methods have also demonstrated the presence of a new clinical form of PCP, termed colonization, which corresponds to the detection of P. jirovecii DNA in respiratory samples in the absence of clinical and radiological features of PCP (see Figure 4). Incidences of colonization have ranged between 9% and 69%, depending on the patient populations studied.86,185 False negatives are however seen when fewer copies are present than the lower limit of detection for a given assay.92 So despite increasing specificity and sensitivity of Pneumocystis detection, PCR interpretation still has challenges in differentiating between active infection and colonization. (See Figure 5 for PCR tracings for Pneumocystis.)
One of the other advantages of qPCR is that it allows for the possibility of quantifying the fungal burden in the respiratory samples.182 Precise cutoffs to distinguish between colonization and infection have not been defined since different techniques and different genes have been targeted in reported studies.186,187 There is also the challenge of a different burden: differences in HIV positive and negative patients. A recent report by Louis and colleagues demonstrated the need for lower cutoffs for HIV negative patients because of lower inoculum size and concluded that different cutoffs must be [End Page 120]
[End Page 121] used in relation to HIV status to differentiate between colonized and infected patients using quantitative real time PCR.182 They used receiver operating characteristic curves (ROC) curve analysis to determine the cutoffs in order to distinguish between colonization and infection, according to the patient's HIV status. It was also observed that prior anti-Pneumocystis therapy to sample collection affected the sensitivity of the test.
β-1-3-D-glucan (BDG). β-1-3-D-glucan is a polysaccharide that is present in the cell wall of most fungi including the Pneumocystis cyst wall.188,189 It has been demonstrated to trigger an intrinsic immune reaction that can be detected in patients' BAL and serum specimens infected with Pneumocystis.189,190 Though BDG specificity for Pneumocystis is relatively poor, it is however highly sensitive for PCP.137,174,191,192 Due to the fact that BDG levels are also increased in other mycotic infections such as candidaemia and aspergillosis, it is better used as an adjuvant test in patients with a high clinical index [End Page 122] for PCP. While doing this is feasible in resource rich countries, it is not cost-effective in resource-poor settings.
In a retrospective study of specimens from 295 HIV infected patients suspected of having PCP, BDG was compared with microscopy with a BDG cut-off level of 31.1 pg/ml. The sensitivity and specificity of the BDG assay were 92.3% and 86.1%, respectively with positive and negative predictive values of 0.610 and 0.980, respectively.174 In this same study, comparison was made with the three other serum markers (CRP, LDH and Krebs von den Lungen-6 antigen (KL-6)) and the ROC suggest that BDG is the most reliable indicator. Another study of 35 PCP patients indicated that, though BDG levels seem a dependable assay for PCP diagnosis, the sensitivity and specificity in patients who were not infected with HIV was lower than in those with HIV infection.193 This could be attributed to the greater burden of Pneumocystis in HIV-infected patients samples compared to that of non-HIV patients. A more recent comparison of the four serological biomarkers (BDG, KL-6, lactate dehydrogenase (LDH) and S-adenosyl methionine (SAM/Adomet)) demonstrated that BDG was the most reliable serologic biomarker for PCP diagnosis, and that the BDG/KL-6 combination test was the most precise serologic assay for PCP diagnosis, with 94.3% sensitivity and 89.6% specificity.194 The BDG/KL-6 combo may provide a less difficult procedure for PCP diagnosis, as it uses blood, which may be also be a significant advantage especially for pediatric cases thus avoiding the associated risk of complications of bronchoscopy. Another study involving 28 HIV-infected patients and 28 control patients with a BGD cutoff value of 100pg/ml revealed a sensitivity of and specificity of 100 and 96.4% respectively.195 More recent studies have revealed that BDG is a useful marker to differentiate between PCP and Pneumocystis colonization. A positive BDG assay result might be a good indication to begin anti-PCP treatment.196–198
S-Adenosyl-L-methionine (AdoMet) plays a vital role in the physiology of all cells, both as a methyl group donor in countless numbers of metabolic processes and as a precursor of polyamines.199,201 It was assumed that P. carinii lacks SAM synthetase so the organism must source this intermediate compound from its mammalian host. However, another study has shown that Pneumocystis possesses a working SAM synthetase.202 Pneumocystis does not synthesize AdoMet, so it scavenges it from the infected host thus suggesting that low plasma AdoMet levels might be a useful marker for PCP.200,201 Several studies from the United States demonstrated that low AdoMet levels could be used to differentiate between PCP and non-PCP pneumonia in HIV-infected patients and healthy control subjects and, additionally, that increasing levels correlated with clinical improvement.200,201,203 One of these studies revealed that Pneumocystis infected patients had considerably lower plasma AdoMet levels when compared with those with non-PCP pneumonia, and no overlap in AdoMet levels were observed between the two groups.201 These researchers found that reduced AdoMet levels were indicative of a PCP infection and when these levels increased it correlated with clinical improvement.201,204 Contrarily, another study measuring serum AdoMet levels found overlaps in levels between PCP and non-PCP pneumonia in HIV infected patients.205 A similar study by another group of researchers did not find any association between AdoMet levels and PCP.206 Whether these differing results relate to variances between plasma and serum AdoMet levels or to other issues is uncertain. Therefore, [End Page 123] more studies are needed to further elucidate the association between Ado Met levels and PCP. Presently, the test is not recommended for clinical use.
A number of factors such as late presentation, missed opportunities for giving prophylaxis (i.e., cough attributed to TB or chest infection, poor compliance and/or side effects), failing ARV therapy207,209 and probably cotrimoxazole resistance inP. jirovecii are all responsible for prophylaxis failures.210,212 Cotrimoxazole, the first line drug for PCP, is widely available and relatively cheap. However, in patients who fail treatment or develop hypersensitivity, drug intolerance or overt toxicity, the alternative therapeutic agents are pentamidine, atovaquone, trimethoprim plus dapsone and clindamycin plus primaquine. All these choices are more expensive and not readily available in resource-poor settings. Prophylaxis. Primary prophylaxis is pertinent for groups such as adolescents, adults and pregnant people with HIV and low CD4 counts or a history of oral candidiasis.10 Guidelines of the World Health Organization (WHO) recommend prophylaxis for HIV-infected people with CD4 counts lower than 350 cells/μL and to consider discontinuation when CD4 counts rise above 350 cells/μL. These guidelines go on to recommend prophylaxis in settings with a high burden of infectious disease or with limited laboratory infrastructure. This position is supported by a recent systematic review and meta-analysis that report benefits of Cotrimoxazole prophylaxis use in HIV positive people (irrespective of CD4 count) in settings with a high burden of infectious diseases.213 The Infectious Diseases Society of America (IDSA) guidelines advocate the use of prophylaxis among human immunodeficiency virus (HIV) infected adults with CD4+ counts less than 200 cells/µl with discontinuation at CD4 counts above 200 cells/µl (Table 4).214 A slightly divergent view supports the discontinuation of prophylaxis earlier among HIV-infected people with CD4 count between 100–200 cells/µl.214 Secondary prophylaxis needs to be initiated in people who have previously had PCP, with continuation until the CD4 count has been above 200 cells/µl for a sustained period. Data from a 12-cohort collaboration suggests that, in HIV-infected people with CD4+ counts of 100–200 cells/mm3 and HIV-RNA levels less than 400 copies/mL, where PCP incidence is low irrespective of PCP prophylaxis use, suggests that it may be safe to stop prophylaxis earlier, however additional data are needed.215
For primary and secondary prophylaxis against PCP, cotrimoxazole , a synergistic combination of the dihydrofolate-reductase inhibitor trimethoprim (TMP) and sulfa-methoxazole (SMX) is the recommended first-line drug.10 TMP/SMX also has the added benefits of being very effective in preventing malaria and malaria-related complications, toxoplasmosis, non-typhoid salmonellosis and many respiratory bacterial infections, and this is particularly beneficial in the tropics that are resource-poor settings.216 A double-strength tablet daily is the preferred regimen but a double-strength tablet given thrice weekly is also effective.10
The frequency of adverse events with use is related to prolonged use of the higher doses.214 Fever, skin rash (Figure 6), nausea, headache and bone marrow suppression are some of the adverse events more commonly documented with cotrimoxazole use. Nephritis and liver function abnormalities are rare events but occur. If adverse reactions are mild, symptomatic management is advocated whilst the patient is encouraged to continue with the drug.10 However, for more serious adverse reactions, it is necessary to change therapy (Table 4). Studies have shown that approximately 25% of [End Page 124] HIV infected patients are unable to endure a full course (21 days) of cotrimoxazole.10 Alternative prophylactic regimens for those who are unable to tolerate cotrimoxazole are dapsone, dapsone plus pyrimethamine plus leucovorin, aerosolized pentamidine and atovaquone.217,218
Factors which influence the outcome of treatment of PCP include the degree of hypoxia at the onset of therapy, the degree of immunosuppression, co-morbid conditions, and tolerability of the most effective agents.219,220 Intravenous formulations are preferred for hospitalized patients, if available. Milder forms of disease are treated with oral therapy from the outset. Treatment is usually limited due to the toxicity of the various agents available.221 For mild to moderate cases of PCP, cotrimoxazole remains the gold standard and first line therapy. The recommended dose is TMP 15–20 mg/kg/day and SMX 75–100 mg/kg/day, given per oral in 3 divided doses10, a total of 120mg/Kg/day initially, given 6 hourly. Cotrimoxazole should be dose adjusted for abnormal renal function and laboratory monitoring of renal function and electrolytes is vital. In pregnancy, cotrimoxazole is the preferred initial therapy in spite of the association of trimethoprim use in the first trimester with an increased risk of neural tube defects and cardiovascular, urinary tract, and multiple anomalies.222,223
The second line drugs include intravenous pentamidine 4 mg/kg IV once daily or primaquine 30 mg (base) per oral once daily plus clindamycin administered either intravenously or orally (Table 5).214 All oral alternatives including trimethoprim plus dapsone, and atovaquone are reserved for mild cases.214 Intravenous pentamidine is the most studied drug as an alternative to cotrimoxazole. Pentamidine was thought to be as active as cotrimoxazole; however, the incidence of adverse events, such as nephrotoxicity and dysglycemia, during treatment with pentamidine is higher.133 Pentamidine use appears to be linked to significantly worse outcomes when compared with other treatment regimes. This is reflected in the findings of a recent study involving 1,188 episodes of HIV-associated PCP cases in Copenhagen (Denmark), London (UK) and Milan (Italy).7 Inferior efficacy was identified as the cause of the increased risk of death associated with pentamidine in this study. This increased risk of death was recorded among the patients switched to pentamidine either because of suspected treatment [End Page 125] failure or because of toxicity.7 The aerosolized form of pentamidine has been found to have limited efficacy and is associated with more frequent relapse.10,224
The combination of clindamycin and primaquine appears to be more effective than intravenous pentamidine for the treatment of PCP where there is clinical failure to cotrimoxazole.225,226 Additionally, oral atovaquone has been demonstrated as being as effective as intravenous pentamidine in the treatment of mild to moderate PCP and has notably fewer treatment-limiting adverse events.217,227 Atovaquone, though less effective than cotrimoxazole is better tolerated and is a useful alternative for mild to moderate PCP.228
Clinical parameters to be monitored to assess response to therapy include respiratory rate, arterial oxygenation, and temperature. The median time to initial clinical response to therapy is 4 to 10 days, although deterioration prior to recovery is common.132 Excess IV fluids should be avoided and fluid overload can mimic failure of therapy.
A number of hospitalized PCP patients require mechanical ventilation with an antecedent high in-hospital mortality.229 Data from a 10 year retrospective study in France revealed that PCP accounted for 35.4% of respiratory failure cases among HIV-infected patients admitted to the intensive care unit (ICU).230 In a more recent cohort study by same researchers, the overall prevalence of PCP reduced significantly to 8.6% at ICU admission and of AIDS and this decrease was progressive over a ten-year period.231 Evidence from a case-control study suggests that the use of noninvasive positive pressure ventilation (NPPV) should be considered as a first-line therapeutic [End Page 126] choice for respiratory failure in AIDS patients with severe PCP to reduce the incidence of pneumothorax and improve survival outcome.232 More recent reports support the use of extracorporeal membrane oxygenation (ECMO) for salvage therapy for respiratory and/or circulatory failure refractory to optimal medical treatment.233,234
Challenge of emerging resistance
Recent data has raised concerns about potential anti-pneumocystis drug resistance due to selective pressure.8,235,236 An appreciation of the magnitude of this problem has been limited somewhat by the lack of in vitro culture systems to facilitate direct testing of the organism. Studies have shown a direct relationship between contact with sulfa containing agents and transmutations of the DHPS gene of P. jirovecii, however the association between these alterations and therapeutic failure is yet to be proven.48,237,238 This information is of utmost importance in development of guidelines for clinicians managing PCP patients. Researchers have focused on direct sequencing of genes that code for enzymes which are targeted by anti-Pneumocystis drugs.239
DHFR and DHPS are the enzymatic targets of cotrimoxazole and dapsone.236 Associations between exposure to sulphur drugs and genetic mutations to the DHFR and DHPS genes have been established.8,239 Mutations at amino acids 55 (Thr3Ala) and 57 (Pro3Ser) in the P. jirovecii DHPS gene have been linked to prior use of sulphur drugs.240–242 However, the role of these mutations in promoting prophylaxis or treatment failure is yet still unclear. Several studies have shown an increased risk of treatment failure with cotrimoxazole in patients with DHPS mutations.242,243 HIV-infected patients with P. jirovecii which have DHFR genetic mutations have been reported to have worse clinical outcomes when compared with patients infected with P. jirovecii strains containing wild-type DHPS.240 The contribution of DPHS mutations and drug resistance to worse clinical outcomes remains uncertain as other investigators have failed to show that DHFR mutations are predictors of PCP mortality; instead they highlight low serum albumin and early intensive care admission as stronger predictors of mortality.244
Availability of first line drugs
The use of cotrimoxazole prophylaxis especially among HIV-infected people with severe immune suppression who are antiretroviral therapy (ART) naïve has been associated with reduction in overall mortality of 19–46% in LMICs.245 In spite of the clear benefits of cotrimoxazole prophylaxis and its availability and affordability, the use of cotrimoxazole among HIV-infected people remains suboptimal worldwide.245 A 2010 WHO survey of 38 countries whose national policies was to provide cotrimoxazole to HIV-infected people revealed that only 25 of the 38 had fully implemented this policy.246 A recent study of 23,816 HIV-infected patients in China reports that 12,047 (51%) had never taken cotrimoxazole, whereas 11,769 (49%) had taken the drug within 6 months of antiretroviral (ART) initiation. Of those who reported cotrimoxazole use, 2,252 (19%) did not begin taking the drug at ART initiation.245 Erratic drug supply, drug stock outs and poor knowledge of cotrimoxazole among both health workers and patients have been identified as root causes of the low utilization.247,248 In many LMICs, intravenous forms of cotrimoxazole are not readily available for the treatment of PCP.
The availability of the other agents for prophylaxis and treatment of PCP is poor when compared to cotrimoxazole. This limits the therapeutic options available to physicians [End Page 127] especially among patients who cannot tolerate cotrimoxazole or a lack of response to treatment with cotrimoxazole. The cost implications of providing alternative therapies is considerably greater than that required to provide cotrimoxazole.
The challenge of toxicity
In spite of the overwhelming benefits of cotrimoxazole use for PCP prophylaxis and treatment, the toxicity associated with its use continues to hinder its use. These toxicities have been found to occur more commonly among HIV-positive people when compared to HIV-negative people.249 The toxicities which limit the use of TMP-SMX tends to occur between day 6 and 10 of therapy.250 The toxic effects include headache, nausea, vomiting, fever, rash, pancytopenia, aseptic meningitis, hepatitis, hyperkalemia and renal dysfunction as seen in Table 5. The life threatening toxicities include a distributive shock syndrome and Stevens-Johnson syndrome.250 On re-introduction of cotrimoxazole following discontinuation due to adverse events such as fever and vomiting, it is better tolerated if the dose is gradually increased (desensitization) according to published regimens.251 It has been reported that up to 70% of patients tolerate such re-institution of therapy.252
Patients on treatment for PCP typically worsen clinically after 2–3 days of starting therapy and this is presumably due to increased inflammation in the lungs as organisms are killed. This worsening is also reflected in an increase in the alveolar-arterial oxygen gradient and can be prevented or blunted with the concomitant administration of corticosteroids at the initiation of therapy. The administration of corticosteroids in conjunction with anti-Pneumocystis therapy has been clearly shown to reduce the incidence of mortality and respiratory failure associated with severe PCP.253 Despite the clear benefits of corticosteroid use in conjunction with cotrimoxazole for moderate and severe cases of PCP in AIDS patients, there are concerns that corticosteroid use increases the risk of opportunistic disease such as cytomegalovirus (CMV), herpes simplex virus (HSV) infections, mycobacterial and fungal diseases, and Kaposi's sarcoma.254,255 Additionally, there is no consensus on the ideal dose and duration of corticosteroid use in PCP treatment. Currently, the WHO does not recommend corticosteroids for PCP in AIDS.
Immune reconstitution inflammatory syndrome and PCP
Despite the undeniable benefits of ART in the setting of HIV, the danger of immune reconstitution inflammatory syndrome (IRIS) in relation to PCP antigens needs to be borne in mind. IRIS can be as a result of immune recovery following ART use thereby unmasking an underlying infection, tumour or disease.256 The two forms of PCP-IRIS are unmasking PCP which occurs within weeks of commencing ART or paradoxical worsening of PCP following stoppage of anti-PCP treatment and initiation of ART.257–260 Cotrimoxazole-resistant PCP remains a strong differential diagnosis of PCP-IRIS and should always be borne in mind. Immune system recovery following ART use in HIV-infected people with PCP favours an effective but exaggerated inflammatory response which is CD4–cell driven.256 This exaggerated response paradoxically leads to severe lung injury and severe PCP. The opposite of this is a CD8+ driven response which occurs in advanced HIV. This CD8+ driven response results in lung damage, a failure to clear the organism and prolonged inflammation which may ultimately be fatal if untreated.256 The optimal timing of introduction of ART in patients with PCP is still open to debate. However, caution is required when introducing ART early in patients with PCP, especially when [End Page 128] adjunctive steroids are used. A case series has described acute respiratory failure occurring after early introduction of ART in patients treated for severe PCP.258 In this study, ART was introduced 1 to 16 days after PCP diagnosis and steroids were stopped on day 15. These patients developed a second episode of severe acute respiratory failure 7 to 17 days after commencement of ART. They all recovered after discontinuation of ART or steroid reintroduction. The marked reduction in plasma viral load recorded in these patients was not matched by an increase in circulating CD4+ cell counts.258
Another case series described the development of a pneumonic illness in people who started ART shortly after PCP had been effectively treated.259 Interestingly, their CD4+ counts had all risen and there was no evidence of PCP, mycobacteria and viruses in the BAL preparations thereby supporting an immune reconstitution illness. This illness could be temporally related to the recovery of their immune system. Therefore, the decision of when to commence ART in HIV-infected people with PCP is further compounded by the fact that clearance of P. jirovecii can be prolonged in HIV-infected people despite clinical improvement.260 A study using direct fluorescent antigen testing demonstrated P. jirovecii cysts in 24% of HIV-infected people 3 weeks after receiving standard PCP treatment.261
Pneumocystis pneumonia still poses a major challenge in managing HIV-infected patients and a high index of suspicion is necessary when caring for this group of patients. Unfortunately, the facilities to diagnose this disease are lacking in many resource-poor settings that are also afflicted by a high burden of HIV cases. Though PCR-based detection techniques are now readily available in many institutions in resource rich countries, they are expensive, require reliable electricity and frozen reagents and thus are not feasible for resource-poor settings. Getting an appropriate sample for testing is challenging especially in children and the lack of bronchoscopy in resource-poor settings remains a limiting factor.
Cotrimoxazole remains the drug of choice both for prophylaxis and treatment though alternative medications are available for patients who are intolerant to cotrimoxazole or in cases of treatment failure. However, there is a dearth of these alternative medications in many resource-poor countries and this might be related to their comparatively higher cost. There is a dire need for cheaper generics and for further drug development to expand the existing armamentarium.
Colonization is a major challenge which needs to be further explored to determine if there is need for therapeutic intervention or if there is a place for respiratory isolation in critically ill AIDS patient.
RITA O. OLADELE is affiliated with the Faculty of Biology, Medicine and Health at the University of Manchester in Manchester, UK and the Department of Medical Microbiology at the College of Medicine of the University of Lagos in Lagos, Nigeria. AKANINYENE A. OTU is affiliated with the Department of Internal Medicine at the College of Medical Sciences of the University of Calabar in Calabar, Nigeria and National Aspergillus Centre at Manchester University NHS Foundation Trust, Manchester, UK. MALCOLM D. RICHARDSON is affiliated with the Faculty of Biology, Medicine and Health at the University of Manchester and the Mycology Reference Centre Manchester at Manchester University NHS Foundation Trust, both in Manchester, UK. DAVID W. DENNING is affiliated with the Faculty of Biology, Medicine and Health at the University of Manchester and the National Aspergillus Centre at Manchester University NHS Foundation Trust, both in Manchester, UK.