Recognition, clinical diagnosis and management of patients with primary antibody deficiencies: a systematic review

Abstract

OTHER ARTICLE PUBLISHED IN THIS MINI-REVIEW SERIES ON IMMUNODEFICIENCY Molecular defects in common variable immunodeficiency. Clin Exp Immunol 2007; 149: 10.1111/j.1365-2249.2007.03461.x

The primary purpose of this systematic review was to produce an evidence-based review of the literature as a means of informing current clinical practice in the recognition, diagnosis and management of patients with suspected primary antibody deficiency. Randomized controlled trials (RCTs) were identified from a search of MEDLINE, EMBASE, The Cochrane Library, DARE (CRD website) and CINAHL by combining the search strategies with The Cochrane Collaboration's validated RCT filter. In addition, other types of studies were identified in a separate search of MEDLINE and EMBASE. Patients at any age with recurrent infections, especially in the upper and lower respiratory tracts, should be investigated for possible antibody deficiency. Replacement therapy with immunoglobulin in primary antibody deficiencies increases life expectancy and reduces infection frequency and severity. Higher doses of immunoglobulin are associated with reduced infection frequency. Late diagnosis and delayed institution of immunoglobulin replacement therapy results in increased morbidity and mortality. A wide variety of organ-specific complications can occur in primary antibody deficiency syndromes, including respiratory, gastroenterological, hepatic, haematological, neurological, rheumatological and cutaneous. There is an increased risk of malignancy. Some of these complications appear to be related to diagnostic delay and inadequate therapy. High-quality controlled trial data on the therapy of these complications is generally lacking. The present study has identified a number of key areas for further research, but RCT data, while desirable, is not always obtained easily for rare conditions. Few data from registries or large case-series have been published in the past 5 years and a greater focus on international collaboration and pooling of data is needed.

Introduction

The primary antibody deficiency syndromes are a heterogeneous group of disorders in which the fundamental defect is an inability to produce an effective antibody response to pathogen. The defects may arise due to disease-causing mutations in single genes of critical importance for the development and function of B lymphocytes. More commonly, they may result from probable environmental triggering in genetically predisposed individuals and are polygenic disorders analogous to diabetes syndromes. These diseases should be distinguished from those disorders where B lymphocytes are affected by a malignant process, as in myeloma or lymphoma [1], those where protein is lost from the gut or urinary tract, or synthesis is suppressed as a consequence of immunosuppressive therapy [2].

The complexity of the immune system is reflected in a huge potential for genetic errors, of varying degrees of significance. In the past 20 years more than 150 primary immunodeficiency syndromes have been identified, and more than 100 of these have been defined at a molecular level, with more genetic defects being identified all the time. Many of these involve defects in antibody production [3]. The main antibody deficiency syndromes of clinical significance are listed in Table 1.

Table 1.

Major types of clinically significant primary antibody deficiency.

Disease	B cells	Immunoglobulins	Inheritance	Gene defect	Mechanism
XLA	Absent/low	Low IgG, IgA and IgM	X-linked	Btk	B cell differentiation failure
CVIDs	≥ 1%	Low IgG and IgA with normal/raised IgM	None	Unknown: homozygous defects e.g. ICOS, TACI Heterozygosity/ polymorphism, e.g. TACI	B cell function failure
Hyper IGM syndromes	Normal	Raised IgM	X-linked	CD40L	T cell co-operation failure
		Low IgG and IgA	AR	CD40
			AR	AID
			AR	UNG
Selective IgG subclass deficiency with/without IgA deficiency	Normal	Normal	Unknown	Unknown	Unknown
Specific antibody deficiencies	Normal	Normal	Unknown	Unknown	Unknown

XLA: X-linked agammaglobulinaemia; CVID: common variable immunodeficiency; ICOS: inducible co-stimulator; TACI: transmembrane activator and calcium-modulating cyclophilin ligand interactor.

In the absence of early diagnosis and appropriate management, primary antibody deficiency can lead to serious morbidity and early mortality. Many complications, both infective and non-infective, can occur. The comparative rarity of these disorders means that levels of awareness outside the immunology community are low, often resulting in delay in diagnosis. Although many of the single gene disorders are identified during infancy or childhood, the most frequent forms of primary antibody deficiency – common variable immunodeficiency disorders (CVIDs) – do not usually manifest until adolescence or early adulthood [4,5], and may not even declare themselves until well into middle age or later. Increased awareness of the possibility of a new diagnosis of primary antibody deficiency in an individual patient is therefore needed for all age groups.

Two main groups have developed diagnostic criteria: the European Society for Immunodeficiencies (ESID) and the Pan-American Group of Immunodeficiency Societies (PAGID). These have been developed based on clinical presentation with genetic analysis, where applicable, but no validation studies have been published. The above diagnostic criteria now serve as case definitions. This systematic review focuses on the clinical aspects of primary antibody deficiencies, including management and complications of primary antibody deficiencies in particular. Systematic reviews aim to be more explicit and less biased in their approach to reviewing a subject than traditional literature reviews, and may provide more transparent evidence for clinical decisions and guidelines and policy based on an appraisal of the wider literature. The primary purpose of this study was to provide validated support to answer a series of questions often posed by physicians and paediatricians. These are given as headings in the main text, followed by the evidence available, in addition to tabulated summaries as appropriate. This was achieved by describing the numbers and characteristics of the relevant studies identified based on a systematic literature search in relation to the clinical features of primary antibody deficiency and indications for therapy. Quantitative analysis (meta-analysis) was not a stated main aim for this review. Molecular aetiology and laboratory diagnosis are covered in detail elsewhere and are beyond the scope of the current review [6].

In order to facilitate the main focus of this review, the individual diseases are not considered separately, as they are too uncommon. All primary antibody deficiency diseases are considered as a group for the purposes of treatment outcomes although the complications of each type of disease are reviewed separately where this is appropriate.

Methodology

Searching

Two main strategies were employed. As part of a report on the effectiveness and cost-replacement of immunoglobulin replacement therapy by the West Midlands Health Technology Assessment Collaboration, randomized controlled trials (RCTs) were identified from a search of MEDLINE (1950–April 2005), EMBASE (1974–April 2005), The Cochrane Library (2005, Issue 2), DARE (CRD website) and CINAHL (1982–2005) [7], by combining the search strategies with The Cochrane Collaboration's validated RCT filter. In addition, other types of (non-randomized) studies − including cohort, case–control and observational studies − were identified in a separate search of MEDLINE and EMBASE by combining comparable search strategies with appropriate search filters (full copies of all search strategies can be obtained by contact with authors).

Inclusion/exclusion

Citations and abstracts were screened for relevance to the review topic by one reviewer in the case of the RCT search, and by two reviewers in the case of the non-randomized studies search. Relevance was determined by the condition (primary immunodeficiency), intervention (immunoglobulins and synonyms thereof), as well as the study type. Searches were supplemented by additional articles known to reviewers. Final decisions on inclusion were made by the same reviewers on the basis of the full text, although two additional reviewers also cross-checked the entire search results for non-randomized studies. Case studies of individual cases or a few patients, while often identified, were not included in the final analysis. Full copies of all relevant papers were circulated to all group members.

Analysis

The primary purpose of this review was to produce an evidence-based review of the literature as a means of informing current clinical practice in the ‘recognition, diagnosis and management of patients with suspected primary antibody deficiency’. Each section was allocated to two reviewers who together reviewed all the relevant literature. Individual publications were rated as recommended by the Scottish Intercollegiate Guidelines Network (SIGN, March 2004 − see Table 2).

Table 2.

Levels of evidence.

1 + +	High-quality meta-analyses, systematic reviews of randomized controlled trials (RCTs), or RCTs with a very low risk of bias
1 +	Well-conducted meta-analyses, systematic reviews of RCTs, or RCTs with a very low risk of bias
1-	Meta-analyses, systematic reviews of RCTs, or RCTs with a high risk of bias
2 + +	High-quality systematic reviews of case–control or cohort studies
	High-quality case–control or cohort studies with a very low risk of confounding bias, or chance and a high probability that the relationship is causal
2 +	Well-conducted case–control or cohort studies with a very low risk of confounding bias, or chance and a moderate probability that the relationship is causal
2-	Case–control or cohort studies with a high risk of confounding bias, or chance and a significant risk that the relationship is not causal
3	Non-analytical studies, e.g. case reports, case-series
4	Expert opinion

Results

The initial search for non-randomized studies was undertaken on 16 August 2005; this was updated subsequently to 13 June 2006. The search strategies for all years (pre-August 2005) identified 408 references; the updated search identified an additional 58 references. A majority of excluded references were case studies of individual cases or a few patients and were not included in the analysis. Full text copies of 120 papers were distributed to all group members. The search for randomized trials was undertaken on May 2005 and identified 891 initial hits yielding 17 randomized trials, the majority of which were not directly relevant, being focused on comparing different immunoglobulin formulations. The search for additional randomized trials was undertaken on 13 June 2006 and identified 152 references. The structure of this report is based upon the following section headings covering referral, management and complications.

1.Who should be referred for specialist assessment?

Clinical history is the most important aspect of suspecting a diagnosis of primary antibody deficiency. Patients at any age with recurrent infections, especially in the upper and lower respiratory tracts, where the frequency or severity of infection is unusual or out of context, for example non-smoking patients, should be investigated for possible antibody deficiency [4,5,8]. There are a number of clinical clues that should raise suspicions of immune deficiency in adults and children, based on national or local registry data or larger case-series [9–15]. Data for presenting symptoms of infection in patients with primary antibody deficiencies are summarized in Table 3, based on six larger reports, in addition to data from the UK audit [16]. Variability between the studies may reflect differences between definitions and patient groups.

Table 3.

Presenting symptoms of infection in cohorts of patients with primary antibody deficiencies.

Site of infection	[12]	[5]	[4]	[8]	[9]	[10]	[16]
Respiratory/chest infections includingpneumonia (excluding ‘bronchitis’)	66%	77%	58%	37%	69%	90%	67.5%
Recurrent sinus infections	60%	98%*	38%	19%	80%	66%	23.3%
GI infections	17%	6%	12%	7%	19%	38%	10.8%
Cutaneous infections	5%	1%	19%		13%		7.9%
CNS/meningitis	6%	2%	6%	2%	9%	4%	4.4%
Septic arthritis/osteomyelitis	1%	2%	7%	2%	1%		1.2%
Ophthalmic	10%		2%				1.4%

^*Includes bronchitis, sinusitis and otitis media. GI: gastrointestinal; CNS: central nervous system.

Not all patients present with recurrent infections, and more rarely autoimmune features are the initial manifestation of primary antibody deficiency. Many patients, especially those with CVID, develop autoimmune problems during the course of their illness [14,17–19]. Other clinical presentations, which require a low index of suspicion for primary antibody deficiency, include abnormal lymphoid tissue, such as nodular lymphoid hyperplasia in the gut or the congenital absence of tonsils, in addition to unexplained signs such as hepatosplenomegaly or arthropathy. Reports in the literature of other non-infectious complications include atopic skin disorders and allergy, haematological complications, organ-specific autoimmune diseases such as thyroid disorders and diabetes, unspecified arthropathy and connective tissue disorders [4,5,10,12,14].

When to suspect primary antibody deficiency in children

The same principles apply to children as to adults, although some differences exist regarding presentations. Children may present with recurrent pyrexia of unknown origin, poor growth and failure to thrive, difficulty coping at school and/or poor school or nursery attendance, usually in conjunction with repeated infections in respiratory and gastrointestinal tracts [14,15,19–21]. Hausser et al. [14] reported short stature (presumed secondary to chronic illness) in seven of 30 children with primary immunodeficiency disorders. Single gene disorders are responsible for primary antibody deficiency in a higher proportion of patients who present in childhood than those who present later in life.

Diagnostic delay and its consequences

The prevalence of clinically significant primary antibody deficiency (PAD) is around 1:25 000 population [19,20,22–24]. Based on United Kingdom workforce data, it is likely that only one in four consultant physicians and one in 15 general practitioners will see a patient with a CVID, the most common PAD, during their working life.

A UK national audit revealed that patients present to a wide variety of clinical specialities other than clinical immunology, but the diagnosis of antibody deficiency was established in far fewer patients seen in these departments than in patients seen in immunology clinics [16]. Both the UK audit [16] and Seymour et al. [8] indicated that respiratory physicians are the most common source of referrals, also showing that there is still a considerable delay in diagnosis of primary antibody deficiency, with a median delay of 1 year. This figure for diagnostic delay represents a small reduction in diagnostic delay from a median of 3.5 years [16], although the main reasons for this reduction are not clear. In the United Kingdom, it may be attributable to the publication of previous reports, including the UK Consensus Document in 1994 [25] and the distribution of national guidelines in 1995. The delay in diagnosis depended on the type of antibody deficiency, and for some conditions (e.g. IgG subclass deficiency) the average delay was in excess of 10 years [8].

The reasons for diagnostic delay are unclear, but plausible explanations include the following:

• Recurrent infections may be accepted as variations of normality rather than investigated.

• There may be a perception that primary antibody deficiency presents exclusively in childhood. However, 95% of patients with primary antibody deficiencies present after the age of 6 years, with a peak presentation in the third and fourth decades of life [5].

• Recurrent infections and the other presenting features of PADs can be difficult to distinguish from other underlying conditions and may be mistaken for other more familiar conditions, such as sarcoidosis or bronchiectasis, or secondary to asthma or chronic obstructive pulmonary disease (COPD).

Although diagnostic delays have decreased in the last 20 years [8,16,26], late diagnosis remains a significant problem. These delays are associated with considerable morbidity, particularly recurrent pneumonias resulting in structural lung damage such as bronchiectasis, pulmonary hypertension and ultimately cor pulmonale [4,8,11,21,26–32].

2.Management of primary antibody deficiency

Immunoglobulin replacement therapy − Table 4

Table 4.

Immunoglobulin replacement therapy in primary antibody deficiency.

Issue	References	Grade of evidence
Increased life expectancy	[5]	2 + +
	[33]	3
Reduction in the rate of bacterial infection	[28,34,35]	2 + +
	[29,36]	2 +
	[37]	2 −
	[38]	3
Diagnostic and treatment delay results in greater morbidity	[4,8,11,28,32]	2 + +
	[21,26,29]	2 +
	[30,31]	3
Risks of therapy	[34]	1 + +
	[46,47,49]	2 + +
	[92,93]	2 +
	[48,94]	2
	[88,95]	3
Reduction in infections with increased dose	[51,52]	1 + +
	[11,28,32,34]	2 + +
	[36,59]	2 +
	[96]	2
	[37]	2 –
	[38,53,61,97]	3
	[16]	4

There are apparently large numbers of RCTs informing the effectiveness of treatment. Unfortunately, most focus on comparing immunoglobulin preparations. There are no RCTs comparing immunoglobulin with no treatment or placebo. There are RCTs comparing high- and low-dose immunoglobulin, which form an important part of the evidence on effectiveness below, but the majority of the evidence relies on observational research.

What are the benefits of immunoglobulin therapy?

There is limited information on mortality, as only two studies were designed to provide survival data [5,33]. However, it is possible to compare data from the 1950s [33] with that from the 1990s [5]. The Medical Research Council study [33] reported 201 case histories of a mixture of immune deficient patients treated with low doses (0.1 g/kg/month) of intramuscular immunoglobulin; the resulting survival time analysis by Kaplan–Meier plot indicated a 10-year survival of 37%. Cunningham-Rundles and Bodian [5] reported 248 patients with CVIDs, the vast majority of whom received intravenous immunoglobulin treatment (IVIg) (0.4 g/kg/month) throughout the period of observation. Although the expected survival in the general population at 10 year was 97%, the CVID patients had a 10-year survival of 78%. Liu et al. [7] reviewed the data and concluded that although mortality was markedly higher in those treated with low dose immunoglobulin, great care needs to be exercised in automatically attributing the improvement to dose of immunoglobulin alone.

There is also a reduction in the rate of bacterial infections, days of antibiotic usage, days of fever and hospital admissions [28,29,34–38] following immunoglobulin therapy. In early studies, in which diagnostic delay was common, pre-existing morbidity obscured some of the benefits of immunoglobulin therapy. However, with the use of higher doses and less marked diagnostic delay, this benefit has been shown to be statistically significant. More recently a significant reduction in pneumonia incidence [from 0.82 to 0.12 pneumonia/patient/year (P = 0.006)] has been shown for patients starting on replacement immunoglobulin therapy [28]. Immunoglobulin therapy, given by both the intravenous (i.v.) and the subcutaneous (s.c.) routes, is equally efficacious in infection reduction [39].

Diagnostic delay inevitably results in treatment delay and avoidable morbidity following further infections. Diagnostic delay has been shown to be associated with higher morbidity measured by infection scores [7]. An episode of pneumonia before treatment results in a 10-fold increase in risk of pneumonia after therapy [29]. Long diagnostic delay is associated significantly with subsequent reduced lung function (P = 0.007) [30].

What are the risks of immunoglobulin therapy?

Risks are divided into two types: infusion related adverse events and the risks of transmission of blood-borne viruses by the immunoglobulin product. Infusion-related risks of adverse reaction have been reduced considerably in recent years, due to improved manufacturing processes, as demonstrated in several trials of new intravenous products [40–45] in which the rate of infusion-related reactions has gradually decreased, although there is still a large regional variation depending on practice [46].

Viral transmission has not been reported since the last outbreak of hepatitis C (HCV) in 1994 [47–49]. Increased regulatory vigilance means that the lack of evidence truly reflects the absence of new outbreaks. Although there is no evidence that limited annual HCV polymerase chain reaction (PCR) is useful, in the event of a further outbreak previous negative results would be invaluable [50].

Does an increased dose of immunoglobulin improve outcome?

Formal studies comparing different doses of immunoglobulin therapy in PAD are limited [34,51,52]. There is evidence that the frequency of infectious episodes was reduced in the high-dose groups compared with the low-dose groups, although the studies did not show statistical significance. Eijkhout et al. [51] also showed that the total episodes of infections per patient were significantly fewer, and that the duration of infections was significantly shorter, for patients in the high-dose compared with the low-dose groups. They also suggested that fever due to infection, use of antibiotics, hospital admissions, the number of patients admitted to the hospital and events of absence from school or work were reduced, although these were not statistically significant. Extrapolation of data from a large study to compare two intravenous products [34] showed reduced rates of documented infections on doses of 0.4 g/K/month (10.5% in 38 patients) compared with those on < 0.4 g/K/month (14.3% in 35 patients) as well as reduced use of prophylactic antibiotics and days of infection-related fever, although these were not statistically significant either.

Four previous studies comparing IMIg (at 0.08–0.1 g/kg/month) with IVIg (at 0.4 g/kg/month) also showed that higher doses of immunoglobulin reduce the rates of bacterial infections [36,53–55]. Ammann et al. [54] found that the IVIG group had a significant smaller percentage of patients with infections in the upper respiratory tract. Two studies also found fewer days of acute respiratory symptoms, fever due to infection and days of antibiotic usage in the IVIg group compared with those in the IMIg group [36,55]. Data from Ochs et al. [56] cautions that there was no improvement in use of antibiotics and in days of missed school or work with doses of 0.4 g/kg/month in the first 12 months if no loading doses were given. Data to support the contention that, in general, higher trough levels are associated with fewer infections can be extrapolated from these trials. Whether or not this is truly associated with an improved outcome has yet to be established.

Is there a direct relationship between IgG trough level and reduced infections?

Another study, designed to show that patients can self-infuse a given dose safely by the intravenous route [57], demonstrated that trough levels needed to prevent infection in individual patients were variable. In 12 patients stabilized on different IgG trough levels (5.2–10.8 g/l), there were no major infections and only four moderate infections in 144 patient-months. Recently it has been shown that there is a significant correlation between prevention of pneumonia and IgG trough levels (P = 0.012) [28].

Two studies from widely differing regions in Europe (Finland and Italy) have shown that low trough levels of IgG are associated with moderate bronchiectasis [11,32]. Structural damage was present in 15 of 22 (severe in five), fibrosis in 21 of 22 (severe in five) in patients with IgG levels < 6 g/l, although not necessarily in the same patients. Furthermore, significantly higher rates of pneumonia (per patient-year) were found in patients with IgG levels < 5 g/l (P = 0.06) [32]. There was also an increased risk of chronic lung disease/sinusitis with time in patients with low IgG levels due to lower doses of replacement immunoglobulin [11]. In a survey of 90 patients with enteroviral infection, data on trough IgG levels suggested a link between low levels and infection, with higher trough IgG levels contributing to a reduction in reported incidence [58]. Similar data relating to mycoplasma infections support the view that infective complications can be reduced by maintaining an adequate trough IgG level [59], although this level remains undefined. Data from Roifman et al. [34] showed that an IgG trough level > 9 g/l reduced validated infection rates from > 10% to 5.6%. This was true even for suspected plus documented infections, by a decrease from 20.9% in those with IgG < 7 g/l, to 18.6% in patients with IgG levels of 7–9 g/l, to 13.6% when the IgG trough level was > 9 g/l. It is important to note, however, that this does not translate into ideal IgG levels for all patients, as the level at which infections are prevented varies widely between patients.

Adjunctive therapy

There is no evidence available on the effectiveness of adjunctive physiotherapy or antibiotic prophylaxis in primary antibody deficiency. Such studies have not been, and are unlikely to be undertaken, as both the selective use of prophylactic antibiotics and physiotherapy are considered widely to be beneficial.

3. Complications of primary antibody deficiency

Acute infections − Table 5

Table 5.

Common acute infections in primary antibody deficiency.

Infection	Common causative organisms	Reference	Level of evidence
Sinopulmonary	Streptococcus pneumoniae	[4,5,10,14,32,51]	2 + +
	Haemophilus influenzae	[63,65]	2
	Moraxella catarrhalis
	Staphlyococcus aureus
Septic arthritis	Mycoplasma spp.	[59,98]	2 –
Genitourinary	Ureaplasma spp.	[98]	2 –
Meningitis	Streptococcus pneumoniae	[4,5,12]	2 + +
	Haemophilus influenzae
Gastrointestinal	Cryptosporidium spp.	[74]	2
	Giardia spp.	[4,12,14]	2 + +
Meningoencephalitis (in XLA)	Enterovirus	[87]	2-

XLA: X-linked agammaglobulinaemia.

Inadequate replacement therapy with immunoglobulin or delayed diagnosis places the patient at greater risk of local or systemic infections. Some of the more common infections are listed in Table 5.

Chronic disease − Table 6

Table 6.

Summary of organ-specific complications.

Feature	References	Level of evidence
Respiratory
Respiratory symptoms are the most common presenting feature of PAD	[4,5,8,10,12,14,28,29,32,37,51,60–66]	2 ++
The diagnosis of PAD is often delayed despite the presence of chronic respiratory symptoms	[5,8,11,12,29,32,60,62,63,66,68]	2 −
Bronchiectasis and bronchial wall thickening occurs in 17–76% patients	[4,8,10–12,14,37,60,62,63,66–69]	2 + +
Lung disease may occur despite optimal immunoglobulin therapy	[37,65]	2 –
Gastrointestinal
Chronic diarrhoea occurs in 40–60% patients	[4,10,14]	2 + +
Infectious diarrhoea occurs in 5–32% patients	[4,10,12,14]
Villous atrophy is present in 2.5% patients	[4,5,14,71,72]	2 –
Nodular lymphoid hyperplasia found in 0.5–30% patients
Others
Atrophic gastritis
Pernicious anaemia
Inflammatory bowel disease
Hepatic
Infectious hepatitis is rare	[4,5,14]	2 + +
Abnormal liver function tests are common
Hepatomegaly occurs commonly
Primary biliary cirrhosis and sclerosing cholangitis are unusual complications	[74]	2 –
Haematological
ITP, AIHA, Evans and neutropenia occur in 2.5–11% patients	[4,5]	2 + +
Overall haematological complications occur in 30% patients	[5]	2 + +
Malignancy
Risk is increased 1.8–13-fold	[79,80]	2
Lymphomas are the most common	[4,5,80–82]	2
Epithelial malignancy occurs more rarely
Neurological
Encapsulated bacterial infections are the commonest cause of meningitis	[4,5,12,86]	2 + +
Unexplained neurodegenerative disease can occur despite IVIg therapy	[88]	3
Rheumatological
Non-specific arthritides	[90,91]	2–
Association with immune-mediated connective tissue diseases is rare
Cutaneous
Infection	[5]	2 + +
granulomata

PAD: primary antibody deficiency; ITP: idiopathic thrombocytopenia; AIHA: autoimmune haemolytic anaemia; IVIg: intravenous immunoglobulin treatment.

Respiratory complications

Respiratory symptoms including productive cough, wheeze, recurrent respiratory tract infections, chronic bronchitis and rhino-sinusitis are the most common presenting features of PAD [4,5,8,10,12,14,28,29,32,37,51,60–66]. Finger clubbing is uncommon in PAD, but may occur in association with granulomatous lung disease (GLD) or bronchiectasis [67]. End-stage lung disease with the development of cor pulmonale and respiratory insufficiency has been documented as the most common cause of morbidity in this group of patients [4,5], because the diagnosis of PAD is often delayed by several years despite the presence of chronic respiratory symptoms and morbidity [5,8,11,12,29,32,60,62,63,66,68]. Patients with milder symptoms such as gastrointestinal or skin infections are often diagnosed later than individuals presenting with unexplained chronic sinusitis and lower respiratory tract infections [32]. Patients developing chronic respiratory symptoms should be managed jointly with a respiratory physician in a multi-disciplinary team approach, as progression of lung disease may occur despite apparently optimal immunoglobulin therapy [37,65].

Pathogens isolated from sputum

The most common pathogens isolated from the sputum of patients with PAD are non-encapsulated Haemophilus influenzae, Streptococcus pneumoniae and Streptococcus pyogenes, with Pseudomonas aeruginosa and Moxarella catarrhalis occurring less frequently [10,14,32,51,63–66].

Abnormalities in pulmonary function

The prevalence of abnormal pulmonary function tests in patients with PAD is unclear. Cohort studies have reported normal lung function in 45–74% of patients [60,63,67]. Abnormal lung function has been reported to be more common in patients with CVID than X-linked agammaglobulinaemia (XLA) [65]. An obstructive pattern is described most frequently, although restrictive disease is well recognized [11,60,63,66,67]. A reduced rate of carbon monoxide uptake per unit time (kCO) may be seen in patients with both obstructive and restrictive patterns [60,65,66]. GLD is associated more commonly with a restrictive ventilatory pattern and reduced transfer factor [67].

Bronchiectasis

Bronchiectasis and bronchial wall thickening are well-recognized complications of PAD [4,8,10–12,14,37,60,62,63,66–69]. The precise prevalence of bronchiectasis in patients with PAD is unknown. Various cohort studies have reported prevalence rates ranging from between 17% and 76% [4,8,10–12,14,37,60,62,63,66–69] Tubular (cylindrical) bronchiectasis occurs in the majority of cases with varicose bronchiectasis occurring infrequently [68,69]. Bronchiectasis occurs most commonly in the right middle and lower lobes [11,60,68,69]. Chest high-resolution computed tomography (HRCT) should be considered in all patients with chronic chest symptoms to monitor progression of disease, as the chest X-ray remains an insensitive tool for the diagnosis of early lung pathology. Other pathologies, including pyogenic lung abscess, empyema and cryptogenic organizing pneumonia, have been described [29,60,63,67,70].

Granulomatous lung disease (GLD)

This has been well described in PAD and is associated with a worse prognosis. GLD is more commonly associated with dyspnoea, splenomegaly and restrictive lung defect, although the presence of normal pulmonary function tests does not exclude the diagnosis [62,67]. The salient radiological features of GLD include diffuse interstitial infiltrates on plain radiograph and consolidation, ground glass and reticular abnormalities on HRCT [62,67].

Gastroenterological complications

Gastrointestinal disorders are reported in 20–47% of patients with primary antibody deficiency [5,12,22]. Most of the reported case-series describe a range of specific gastroenterological disorders. However, the number of cases requiring histological diagnosis is rarely stated as a proportion of those actually investigated by endoscopy. Given that the threshold for performing endoscopic evaluation varies, and as many patients with gastrointestinal symptoms remain uninvestigated, the true incidences are not known.

Chronic diarrhoea

Chronic diarrhoea of unspecified cause occurs commonly in both CVID and XLA, frequencies ranging from 40% to 60% [4,10,14]. This includes the complications listed in Table 6.

Infectious diarrhoea

The most frequently identified organisms are Giardia lamblia, Campylobacter spp. and Salmonella spp., all of which occur with increased frequency in untreated patients with both CVID and XLA. The overall reported incidence of infectious diarrhoea varies from 5% to 32%. The highest incidence is reported in Hermans et al. [10], possibly reflecting later recognition and less ideal therapy than in more recent series [4,12,14].

Villous atrophy

Small bowel villous atrophy can be found in those patients investigated by endoscopy, occurring at an overall frequency of around 2.5% in cohort studies [4,5], and possibly higher in children [14]. In case-series investigating patients with gastrointestinal symptoms, the incidence is higher [71,72]. A clinical response to exclusion of gluten from the diet is reported in some cases, although it is rarely clear whether repeat biopsies have been performed. In most cases no response to gluten withdrawal is observed or is not reported to have occurred.

Atrophic gastritis and pernicious anaemia

Atrophic gastritis is reported in patients with CVID, but not XLA, and in some cases is associated with pernicious anaemia. True incidences cannot be determined from the available studies [4,10,12,14].

Nodular lymphoid hyperplasia

Hyperplasia of gut lymphoid tissue is seen frequently in patients with CVID. True incidences are difficult to estimate because of variation in frequency of endoscopic evaluation. The data available suggest incidences of between 0.5 and 30% [4,5,14,71,72]. It is not possible to determine whether or not this is infective in origin.

Inflammatory bowel disease (IBD)

Inflammation of both small and large bowel is reported in 2–13% of patients with PAD. Various types occur, including autoimmune, Crohn's-like and non-specific [4,5,12,14,71,73].

Hepatic complications

Various liver disorders are observed in PAD. Hepatitis C has been associated with infusion of contaminated batches of immunoglobulin in the past (see section 2), but the risk has been very low since the introduction of more stringent manufacturing processes and screening procedures. Other forms of infectious hepatitis can occur, although this is not a frequent complication [4,14]. Abnormal liver function is common, and a clear cause may not be definable [5]. Hepatomegaly may be caused by granulomatous infiltration, and is often associated with splenomegaly [4]. Other disorders that have been associated include primary biliary cirrhosis [5] and sclerosing cholangitis, particularly in children, and often associated with infection with Cryptosporidium parvum[74], a particular complication of combined T and B cell deficiency, as in CD40 ligand deficiency.

Haematological complications

A variety of haematological disorders are recognized in association with PAD, and have been reported in many case-series. Most frequent are autoimmune cytopenias, which may occur as the initial presenting feature of immunodeficiency (see earlier section).

In the largest case-series of CVID reported (326 patients) [17], 35 (11%) had a history of autoimmune haematological disease. Of particular note is that 30 of these (86%) had their first episode of thrombocytopenia or haemolytic anaemia before, or concurrent with, the diagnosis of CVID. Broadly comparable numbers of patients with idiopathic thrombocytopenia (ITP), autoimmune haemolytic anaemia (AIHA) and Evans' syndrome (both ITP and AIHA) were reported (15, nine and 11 patients, respectively). In patients who had ITP or Evans' syndrome, the age at diagnosis of CVID was younger (50% diagnosed by age 25 years) than in those who developed AIHA (50% diagnosed by age 50 years). The association between ITP and CVID has also been reported by other groups, including Michel et al. [75], who emphasize that in the majority of patients ITP occurs before the diagnosis of CVID.

An earlier report of a subset of the above large case-series [5] reported similar incidences of AIHA and ITP, but additionally a small number of patients with autoimmune neutropenia (two of 248: 0.8%). The largest UK case-series [4] reported haematological complications overall in 94 of 240 (39%) patients with CVID, and eight of 44 (18%) with XLA. Lymphopenia was included and was common, occurring in 50 of 240 (21%) CVID, and five of 44 (11%) XLA. Twelve of 240 (5%) patients with CVID had AIHA, and in six (2.5%) of these the AIHA occurred before the diagnosis of immunodeficiency. In this series, ITP was less frequent (six of 240, 2.5%, in patients with CVID). Neutropenia was seen in 21 patients, the majority of whom suffered from XLA (18 of 44; 41%), with only three of 240 (1.25%) in CVID. Other case-series and national registries have reported variable numbers of autoimmune cytopenias in patients with PAD [12–14,20,22,76].

Malignancy

The risk of malignancy in patients with primary antibody deficiency is well-recognized [77], estimates of the overall increased risk, varying from 1.8-fold [78] and fivefold [79] to 13-fold [80], and appears to be linked with the primary antibody deficiency rather than the genetic background of the individual [78]. The risk of cancer in patients with CVID varies with age, from 2.5% in the under 16-year-olds to 8.5% in the 16 + age group [81]. Lymphoreticular malignancies are most frequent, with non-Hodgkin's lymphoma predominating, particularly in females [4,5,80–82] Lymphomas are predominantly B cell in origin [83,84]. Epithelial malignancies also occur with increased frequency, particularly gastric carcinoma [10,22,79]. Cancers in many other organs are also reported [4,5,10,12,76,81,85]. Although the risk is particularly high in CVID, there is also a significant increase in other forms of primary antibody deficiency [81,85].

Many case-series include patients with unexplained lymphadenopathy and/or splenomegaly [4,12]. Some of these have granulomatous changes, and some have ill-defined lymphoproliferative disorders, both of which can also occur and co-exist in other organs. Epstein–Barr virus is found in association in only a minority of patients with lymphoproliferative disease (LPD). Monoclonal or oligoclonal populations of lymphocytes may be found in lymphoproliferation and lymphoma, and the distinction between these is difficult in patients with PAD [84].

Neurological complications

Infections

Cohort studies [4,5,12,32,86] have reported central nervous system (CNS) infection both as presenting feature and complication of primary antibody deficiency, particularly prediagnosis or during suboptimal therapy. In a study of 73 XLA patients [32], 4% suffered from meningitis before diagnosis. In the cohort studies of CVID [4,5,12,86] meningitis was present prediagnosis in 1–6% of patients. One study [4] reported the causative organisms as S. pneumoniae, Neisseria meningitidis, S. aureus, Escherichia coli, H. influenzae and Listeria monocytogenes, with pneumococcal disease being the most common. Within these studies, other neurological presentations include encephalitis, meningoencephalitis, brain abscess and vaccine-induced poliomyelitis. McKinney and coworkers [87] reported a series of 42 XLA and CVID patients who had contracted enteroviral infection. As a cause of death, neurological complications appear relatively unusual. There are reports of cytomegalovirus (CMV) and measles encephalitis, cerebral atrophy, cerebrovascular accidents and progressive multi-focal leucoencephalopathy (PML) as causes of death.

Neurodegeneration

Ziegner et al. [88] reported 14 patients who had developed an unidentified neurodegenerative condition. There was no consistent diagnostic approach between reporting centres and no details of complications such as malabsorption or enteropathy. Because seven suffered from ataxia and incoordination, four had tremor and five also suffered from retinitis pigmentosa, all known to be associated with vitamin E deficiency, this may have been the cause of symptoms in some of these patients [89]. Two studies [5,86] reported cerebral atrophy in two patients and as a cause of death in one patient.

Rheumatological complications

Rheumatological complications of primary antibody deficiencies are primarily those due to acute or chronic infection, particularly with Mycoplasma spp., that resolves on appropriate antibiotic therapy and is prevented by the institution of immunoglobulin therapy [59]. They also include seronegative arthritis associated with chronic low-grade infection [18], that resolves once adequate immunoglobulin therapy is started [4], and associated immune-mediated connective tissue diseases (which are rare) [4,18,90,91].

Cutaneous complications

The different dermatological manifestations of primary antibody deficiency include infections of the skin, eczema-related disorders, erythroderma, vasculitis and autoimmune diseases including granuloma formation, and skin problems characteristic of specific immunodeficiency syndromes. Skin infections may be fungal, bacterial or viral. In the largest case-series of common variable immunodeficiency, there were a small numbers of patients who presented with associated infections (e.g. herpes zoster), urticaria/angioedema and alopecia [5].

In surveys involving children, immunodeficiencies may present with complications or more severe manifestations of eczema, although eczema is very common in children. Hausser et al. [14] reported clinical observations in 30 children with CVID; and cutaneous infections were reported in 13% of patients, atopy-allergy in 26% patients, alopecia in 7% and psoriasis in 7%. Wang et al. [15], in a retrospective study of children with hypogammaglobulinaemia, reported clinical presentations of atopic dermatitis, atopic rhinitis and asthma in 36% of children.

Skin manifestations are also recognized in association with autoimmune disorders associated with immunodeficiencies, e.g. lupus. Non-infectious granuloma may also occur in the skin of patients with immunodeficiencies although establishing an accurate prevalence is complicated without defining common criteria for biopsy and diagnosis.

Conclusions

The primary antibody deficiency syndromes are a rare group of primary immunodeficiencies where diagnostic delay remains common due to limited awareness of their presenting features. Referral for specialist assessment leads to earlier diagnosis and therapy. As with many other rare disorders, study of these disorders is a minor part of undergraduate and general postgraduate training. Greater education of health-care professionals is required to ensure prompt recognition and referral to specialists in primary immunodeficiencies. Such an approach is likely to improve measures of mortality, morbidity and quality of life in this patient group.

Replacement therapy with immunoglobulin increases life expectancy and reduces infection frequency and severity. Late diagnosis and delayed institution of immunoglobulin replacement therapy results in increased morbidity and mortality. Increased doses of immunoglobulin therapy improve outcome measures with regard to infection frequency and severity, but whether this correlates with outcome remains to be established.

A variety of complications can occur in primary antibody deficiency syndromes, and some of these appear to be related to diagnostic delay and inadequate therapy. However, despite replacement therapy, patients remain at risk of breakthrough bacterial infections with current management plans. Respiratory problems, particularly structural lung damage, are the most common complication and are more common with greater delay in diagnosis and therapy. A large variety of other system complications can occur and some of these may be inherent in the disease pathogenesis, particularly in CVID. Lack of recognition of these unusual complications contributes to morbidity and decreased life expectancy.

These patients often present to haematology and respiratory medicine departments and a low threshold for referral to and investigation by clinical immunologists would improve the ascertainment of those patients with treatable primary antibody deficiency.

The present study has identified a number of key areas for further research, as follows.

• Cohort analyses of existing data to confirm the effect of treatment on mortality because a placebo-controlled study is unethical.

• The optimal dose of replacement immunoglobulin therapy required to improve overall health outcome measures.

• Identification of prognostic markers to allow specific intervention and optimal therapy for subgroups of patients.

• Determination of the causes of progression of lung disease despite apparently optimal immunoglobulin therapy.

• Prospective, multi-centre, microbiological studies to characterize the precise aetiology of infections in this group of patients and to identify ways in which these can be prevented.

• The role of pulmonary function tests and HRCT in monitoring patients on immunoglobulin replacement therapy.

• Assessments of the efficacy of physiotherapy and various antibiotic prophylaxis regimens to ascertain which of these practices lead to improved management and elimination of breakthrough infections.

• The management of other system complications is often based on consensus approaches in the absence of high-quality controlled trial data. Further research is needed in these areas to define optimal management for end-organ complications.

RCT data, while desirable as the highest level of evidence to inform optimal clinical practice, is not always obtained easily for rare conditions such as primary antibody deficiencies. It is also apparent from the present study that few data from registries or large case-series have been published in the past 5 years. A greater focus on international collaboration and pooling of data would contribute to answering some of the remaining questions in this complex group of disorders.

However, where cohort studies and registry data are the basis of effectiveness estimates, a high level of detail is required on the exact treatments being given and how they vary over time. Similarly, adjustment for the confounding effect of differences in prognostic factors when making comparisons is also a major challenge.