Infections are a major and common cause of morbimortality in patients with rheumatic and systemic autoimmune diseases (RD&SA).1 They include opportunistic infections among others, defined as those caused by non-pathogenic organisms which become pathogenic when the immune system is altered.2 They are an increasingly severe problem, due to the use of immunosuppressor agents to treat diseases, as well as the increased susceptibility to infection caused by diseases themselves and the comorbidities associated with them.3,4

Pneumocystis jirovecii, which was previously known as Pneumocystis carinii, is a pathogenic, opportunistic and extracellular fungus that belongs to the hemiascomicetes class. It is in the Pneumocystis genus, a complex group that is composed of many species with peculiar characteristics that differentiate them from other fungi.5 Their cytoplasmatic membrane is mainly composed of cholesterol, unlike those of other fungi which have membranes that contain ergosterol. This is why standard antifungal drugs (such as the azoles and polyenes), are ineffective against P. jirovecii, as they act on the ergosterol of the fungal membrane, either directly (amphotericin) or indirectly, inhibiting its synthesis (azoles).6 The new antifungal drugs (echinocandins), nevertheless, interfere in the synthesis of 1,3-β-D glucan, a specific constituent of the membrane of all fungi that is found in high quantities in the cystic form of P. jirovecii, so that they would be a highly interesting member of the therapeutic arsenal.7 Another peculiarity of Pneumocystis that is unique in mycology is their specificity for the selective invasion of a concrete host, so that each mammal is affected by a different genetic variety. Only P. jirovecii is found in humans. Lastly, this fungus does not grow in artificial media, so that diagnosis requires demonstration that it is present in sputum samples obtained by broncoalveolar washing, or in tissue obtained by pulmonary biopsy, using microscopy techniques (preferentially with immunofluorescence), or by techniques involving highly sensitive polymerase chain reactions.8

Although its exact transmission mechanism is unknown, the most widely accepted hypothesis is that it passes from one person to another by the inhalation of particles in the air.8 Exposure to P. jirovecii in humans is highly common, as is shown by the fact that in developed countries more than 80% of children have developed antibodies against P. jirovecii before they are four years old.9

Although the theory used to be that pneumonia caused by P. jirovecii (NPJ) is due to the reactivation of a latent infection in an immunocompromised patient, current data show that it is a de novo infection. Nevertheless, colonisation by P. jirovecii has been shown to exist in certain susceptible populations, although the clinical implications of this finding are unclear.5

Once it has been acquired, P. jirovecii passes through a complicated life-cycle. It has two predominant forms, trophic and cystic, of which the first type represents 90% of the total number of P. jirovecii organisms during infection. Trophic forms are united by interdigitations to the cellular membrane of the type I pneumocyte in the alveolar epithelium, permitting the close unification of both membranes without breaking the alveolar cell or penetrating it. The interaction of P. jirovecii with the pneumocyte and alveolar macrophages triggers a cascade of cellular responses in P. jirovecii itself as well as in the lung cells: P. jirovecii is stimulated to proliferate, at the same time that the alveolar macrophages commence phagocytosis of P. jirovecii and its destruction, and the alveolar cells release proinflammatory cytokines and chemokines that promote the recruitment and activation of neutrophils and T lymphocytes. T CD4+ lymphocytes encharged with coordinating the inflammatory response of the host, recruiting and activating additional immune effecter cells (monocytes and macrophages) which will be responsible for eliminating P. jirovecii. In immunocompetent individuals the infection is eliminated with minimum inflammation and lung damage. However, in immunocompromised hosts a hyperinflammatory response is trigged that is unable to eliminate P. jirovecii but which causes lung damage and affects gas interchange. This causes NPJ, the chief manifestation of this infection.

P. jirovecii has a special tropism for the lungs, and its dissemination in the rest of an infected organism is exceptional. Interstitial pneumonia is the main disease caused by P. jirovecii. The most common symptoms are the appearance of a dry cough, dyspnoea, fever, tachycardia and tachypnea.10 In pulmonary auscultation the presence of fine rales stands out, and in thoracic X-ray there is tenuous bilateral interstitial infiltrate. High resolution computed tomography (HRCT) shows higher sensitivity in detecting NPJ than simple X-ray imaging, showing the characteristic areas of ground glass opacity in peri-hiliar distribution.11 Generally high levels of lactate dehydrogenase (LDH) are detected, with a fall in serum albumin and CD4+ counts <200/mm3. Due to damage in the alveolar epithelium the blood gases are altered, detecting hypoxemia and/or respiratory failure in arterial gasometry.

In immunocompromised patients without HIV, the infection usually runs a more acute course, manifesting with fewer systemic symptoms but with greater respiratory repercussion (a higher level of respiratory failure and tachypnea), and in broncoalveolar washing a lower concentration of organisms is detected, although with a higher neutrophil count5 and, in general, the condition tends to be more severe, with a longer stay in intensive care units and higher rates of mortality.8,12 This higher mortality has been associated with low PaO2/FiO2 coefficients,13 with the need for mechanical ventilation,14 hypoalbuminemia, male sex, advanced age and medical care in private hospitals.15

This infection is often lethal.5 The mortality rate is close to 100% without treatment, and it varies from 5% to 40% in treated patients.10 The rate of mortality due to NPJ in patients without HIV currently stands at from 39.4% to 59.1%, a figure far higher than the 6%–7% recorded in HIV.14

The individuals at the greatest risk of NPJ are patients with HIV, especially those with CD4+ counts <200/mm;3 those who have received organ transplants or haematopoietic cells;5 premature infants who require mechanical ventilation; patients with primary immunodeficiencies that affect T lymphocyte functioning16 and patients who receive oncological chemotherapy or are treated with immunosuppressor drugs.17 The use of antiretroviral drugs to treat HIV infection together with prophylaxis against P. jirovecii following the recommendations of clinical practice guides18 have drastically reduced the incidence of NPJ in developed countries in patients with AIDS, from rates of 70%–80% before implementing these measures, to fewer than one case per 100 persons-year. On the contrary, an increase in incidence has been observed in immunocompromised patients without HIV: rates of incidence vary depending on the disease, but they may reach 44.6 cases per 100,000 patients/year in those who have received transplants, >45 cases per 100,000 patients/year in haematological neoplasias, 53.6 cases per 100,000 patients/year in inflammatory myopathies, 71.9 cases per100,000 patients/year in vasculitis associated with anti-neutrophil cytoplasm antibodies (ANCA) and 93.2 cases per 100,000 patients/year in patients with panarteritis nodosa (PAN).19

Guides for the prophylaxis of NPJ have recently been published for patients with haematological diseases and organ transplant.12 Nevertheless, to date no guide for the prophylaxis of NPJ has been published for patients with RD&SA under immunosuppressor treatment. Several authors have expressed the advisability of having these guides,20,21 as they would help to unify criteria which are currently very varied.22,23

As occurs in other pathologies,12 the risk of NPJ in patients with RD&SA will be determined by the combination of different factors: 1) the disease; 2) the drugs used; and 3) the particular circumstances of each individual. When deciding on the advisability of administering prophylaxis against P. jirovecii, it is also necessary to consider the potential toxicity of the treatment used for this.24 It has been suggested that a "level of risk of NPJ" higher than 3.5% should be the cut-off point for considering that the beneficial effect of prophylaxis is greater than the risk.25 Finally, it has to be taken into account that many of these factors will be modified over time, so that continuous re-evaluation will be required of the risk/benefit during the follow-up of the patients.26

The incidence of NPJ varies notably from some RD&SA to others, and it is higher in granulomatous vasculitis with polyangiitis (VGP), PAN, systemic lupus erythematosus (SLE) and inflammatory myopathies, making an individualised approach necessary.

Different drugs have been tested in the prophylaxis of NPJ: TMP/SMX, pentamidine, atovaquone, dapsone, pyrimethamine and clindamycin.57 The best result was obtained with TMP/SMX, and this is why it is the recommended first-line drug. Daily administration is equally effective as administration three days per week.21,58 When TMP-SMX is contraindicated the recommended second-line is pentamidine, atovaquone or dapsone20 (Table 1).

It is important to take into account the interaction between TMP-SMX and methotrexate, (MTX) as this association may cause severe cytopenia and myelosuppression, even at low doses of MTX over a short time (two days) of treatment with TMP-SMX.59 This effect does not seem to arise with prophylactic doses of TMP-SMX, although it must be taken into account. The main side effects associated with the use of TMP/SMX are listed in Table 2.60,61

The duration of prophylaxis is controversial. In ANCA-associated vasculitis it is suggested that it be discontinued when the immunosuppressant is suspended, when the corticoid dose is lower than 20−10mg of prednisone/day, or when B lymphocyte depletion halts after having used RTX.21 Prior to discontinuation other factors should be evaluated: lymphopenia, leucopoenia, the CD4+ count and the dose of corticoid.

There can be no doubt that the incidence of NPJ is higher in patients with certain RD&SA, often with a disastrous outcome. Nor can there be any doubt as to the efficacy of prophylaxis in preventing it. However, there is no clear definition of the circumstances under which it is reasonable to apply the said prophylaxis. This uncertainty is reflected in the result of different surveys of clinics, where a wide range of criteria were detected.22,23

Green et al.25 set the cut-off point at a risk of NPJ above 3.5% so that the risk/benefit balance is favourable. According to these authors, patients with VGP would be included in this category, while the other RD&SA (inflammatory myopathies, SLE, PAN, scleroderma and RA) should not receive prophylaxis as the balance would be unfavourable. In their study Park et al.24 find an overall number necessary for harm (NNH) of 131 (55–∞), as opposed to a number necessary to prevent a case of NPJ (NNT) of 52 (33–124). Stratifying according to diseases, the NNT in patients with SLE (43 (28–85)) or PAM (3 (1.6–39.4)) was lower than the NNH, which would justify prophylaxis. This is not the case for the other diseases. Nevertheless, the actual situation is far more complex, as the risk of NPJ will not only depend on the type of disease, but also on its association with other predisposing factors. A retrospective review of 21 cases of NPJ in patients with RD&SA over a period of 20 years in an American hospital supports the theory that aetiology here is multifactorial, with the participation of several factors (high doses of corticoids, the use of other immunosuppressor drugs, lymphopenia or the coexistence of interstitial pulmonary disease).62 None of the 21 patients had received prophylaxis.

Different authors propose a very wide range of types of prophylaxis:

• •

  Wolfe et al.21 classify NPJ risk factors into three categories: established, probable and possible (Table 3). They suggest commencing prophylaxis:

  • ◦

    Always, during induction treatment in VGP.

  • ◦

    In other ANCA-associated vasculitis and in PAN: during induction treatment or high dose corticoids, on condition that there is lymphopenia (<500cells/mm3) or a low CD4+ count (<200/mm3).

  • ◦

    In SLE and inflammatory myopathies: when high doses of corticoids are being administered, when there is also lymphopenia or a low CD4+ count and in SLE, during immunosuppressor treatment or severe disease in case of inflammatory myopathy.

  Table 3.

  Risk factors for the development of NPJ in patients with RD&SA. Modified from Wolf et al.21.

  Established	Probable	Possible
  Low T CD4+ lymphocytes	Glucocorticoids	Young age
  Lymphopenia	Cyclophosphamide	Male sex
  	Rituximab	Hispanic descent
  	Methotrexate	Asiatic descent
  	Anti-TNF	Private medical treatment
  	Azathioprine	Interstitial lung disease
  		Caucasian descent
• •

  Ogawa et al.,53 Chew et al.63 and Yale et al.17 base the start of prophylaxis on the corticoid dose given, and they propose commencing it when the dose is above 10−30mg of prednisone.

• •

  Stamp et al.20 administer prophylaxis to all of their patients treated with immunosuppressants and those with a history of NPJ, as well as patients with persistent lymphopenia and CD4+ counts below 200/mm3.

• •

  Demoruelle et al.64 propose reserving prophylaxis for patients with connective tissue pathologies who fulfil two or more of the following criteria:

  • ◦

    Steroids ≥20mg/day for more than four weeks.

  • ◦

    ≥ Two disease-modifying drugs (DMD).

  • ◦

    Total lymphocyte count ≤ 350/mm3.

  • ◦

    Parenchymal lung disease.

• •

  Gupta et al.23 propose, in patients with SLE treated with CFM, administering prophylaxis only to those who have an additional risk factor: severe leucopoenia, severe lymphopenia, high doses of corticoids, hypocomplementemia, active kidney disease or high activity scores (SLEDAI).

• •

  Li et al.27 propose administering prophylaxis to patients with RD&SA who receive high doses of immunosuppressant (such as pulses of methyl-prednisolone) and who have CD4+ counts lower than 250/mm3.

• •

  Sowden et al.55 propose determining CD4+ after one month of immunosuppressant only in patients who fulfil the following three criteria: doses of prednisolone or equivalent above 15mg/day, planned duration of treatment longer than three months, and total lymphocyte counts <600/mm3. In cases with CD4+ counts <200, prophylaxis would commence if the annual risk of P. jirovecii is higher than 9%.

• •

  Mansharamani et al.,54 suggest use a CD4+ count lower than 300/mm3 as the cut-off point to commence prophylaxis.

• •

  Mori and Sugimoto43 administer one tablet of TMP-SMX (80mg TMX plus 400mg SMX) per day during 5–7 days, or two tablets three days a week for one week, to all patients with RA prior to starting treatment.

At the current time there are no data on which to decide which of these guidelines is the most effective in clinical practice, and nor have any studies been designed to find predictive factors for the development of NPJ in patients with RD&SA.65 Taking its severity and the possibility of preventing it into account, and awaiting studies that supply more solid and conclusive data, it would be desirable to have recommendations that help clinicians to identify risk situations and apply prophylaxis more efficiently, as this would doubtless have a highly favourable repercussion in our patients.

This research received no specific aid from public, private or not-for-profit bodies.

The authors have no conflict of interests to declare.