INTRODUCTION

Common Variable Immunodeficiency (CVID) is the most frequent symptomatic antibody deficiency, characterized by low levels of serum immunoglobulins and impaired antibody response. Most cases are sporadic, approximately 20% are familiar, displaying autosomal dominant and autosomal recessive traits (1–4). CVID is a late-onset hypogammaglobulinemia, but it may also occur in infancy (5, 6). The clinical spectrum includes respiratory infections, gastrointestinal diseases, autoimmune diseases, granulomatous manifestations, and cancers (7–11). Long-term administration of immunoglobulins reduces the incidence of infections (12–18). However, despite IgG replacement, CVID patients may still suffer from respiratory infections, caused by encapsulated bacteria, leading to permanent lung damage (8–10).

We have collected information on 224 CVID patients enrolled in a multicenter prospective study in order to provide an updated view of the spectrum of illnesses which occurred at the time of onset and over a mean of 11.5 years of follow-up. Moreover, we gathered information on the effect of long-term immunoglobulin replacement therapy.

METHODS

In 1999, 26 Italian Centers belonging to the Italian Primary Immunodeficiency Network started to submit data for patients with a CVID diagnosis. Diagnosis was made in patients over 2 years of age by approved criteria (2) including low levels of serum IgG, IgA, and/or IgM, antibody deficiency with impaired response to tetanus and pneumococcal antigen immunization, peripheral B cell numbers of >2%, and exclusion of hypogammaglobulinemia due to other primary or secondary immunodeficiencies. Detailed information with personal data, pedigree, date of diagnosis, immunological data and clinical manifestations, and route and dosage of immunoglobulin replacement from diagnosis to the time of enrolment were collected in a semistructured questionnaire filed by one physician per center at the time of enrolment and on a yearly basis up to 2005. All data were processed in a database and sent to the Interuniversity Computing Center (CINECA) responsible for processing and analyzing the data. For each patient, complete blood counts, lymphocyte subsets, chemistries, serum immunoglobulin levels, and cultures tests were performed four times per year. Chest and sinus-computerized tomography (CT) scans were performed every 4 years, and gastrointestinal endoscopy with biopsy every 2 years or when indicated.

Our multicenter study has been designed according to the World Medical Association Declaration of Helsinki ethical principles for medical research involving human subjects.

Statistical Methods

All information was stored, controlled, and analyzed by an integrated system of software facilities. Differences between males and females and of ages classes were tested with Wilcoxon rank-sum tests. Fisher's exact test was used to compare differences in percentages for categorical variables. Student's t test was adopted to compare means for continuous variables. All p values are two-sided and values <0.05 were considered significant. The SAS package (SAS Institute, Cary, NC) was used for the analysis of the data. Follow-up time was calculated as the time between diagnosis and the date of last annual questionnaire send to CINECA. The observed numbers of cancer cases were compared with the expected numbers calculated to yield the Standardized Prevalence Ratio (SPR). Associated 95% confidence intervals (CI) were calculated under the assumption that the observed number of cancer cases followed a Poisson distribution, by use of Byar's approximation. Significantly increased SPRs have lower confidence limits that are above or equal to 1.01. This means that the p value for obtaining the SPR estimate is less than 0.05 under the null hypothesis that the observed number of cancers in the study population is not different from the expected number in the general population. Probabilities of survival after diagnosis of CVID were estimated from Kaplan–Meier life tables.

Fig. 1.
figure 1

Age at diagnosis (panel A) and years of diagnostic delay (panel B).

Full size image

RESULTS

We enrolled 224 CVID patients, 111 males and 113 females. The mean age at the time of diagnosis was26.6 years (range 2–73). Seventy-five patients were younger than 14 years of age. The mean age at the onset of symptoms was 16.9 years (range 2–66). This implicates with a mean diagnostic delay of 8.9 years (in 55% of patients, the diagnosis was made within 5 years from the clinical onset) (Fig. 1). Patients were followed-up for a mean time of 11.5 years (range 3–34).

Clinical Manifestations

The spectrum of diseases associated with CVID includes infections, pulmonary diseases, autoimmune diseases, granulomatous infiltrations, gastrointestinal diseases, and cancers. A minority of CVID patients were asymptomatic at the time of diagnosis (4.6%). This percentage increased during follow-up to 8.6% at the last clinical assessment. There was no difference in the percentage of asymptomatic patients between males and females and between adult and young patients (<14 years of age).

Infections.

Acute and chronic infections were found in the majority of patients at the time of diagnosis (196 patients, 100 females and 96 males) and during follow-up (194 patients, 102 females and 92 males) (Table I). Thus, a similar percentage of patients with any infection at the time of diagnosis (87.5%) and during the follow-up period (86.6%) was found.

Table I. Prevalence of Infections at Diagnosis and During Follow-Up in 224 CVID Patients
Full size table

Respiratory Tract Infections.

Respiratory tract infections were the most prominent clinical problem observed at diagnosis and during follow-up. A significant reduction in the percentage of patients affected by acute pneumonia (p < 0.001) and acute otitis (p < 0.001), but a significant increase in the percentage of patients with chronic sinusitis and CLD (p < 0.001) was observed during follow-up. The longitudinal analysis of patient's individual data demonstrated that 49% had pneumonia at least once prior to diagnosis: after starting IVIG replacement, 35.7% never experienced any further episode of acute pneumonia, while 13.3% continued to have recurrent pneumonia. Seven percent of patients experienced the first episode of pneumonia during the follow-up, despite on IVIG therapy (Fig. 2, panel A). About 39% of patients had acute otitis prior to diagnosis. Since the time of IVIG replacement, 25.5% of patients never experienced any further episode, while 13.3% continued to have recurrent acute otitis. Despite IVIG treatment, 12% had the first episodes of acute otitis during follow-up (Fig. 2, panel B). Thirty-six percent of patients were affected by chronic sinusitis at the time of diagnosis. This percentage increased to 54% during follow-up (Fig. 3, panel B). A group (8.6%) of patients who had chronic sinusitis at the time of diagnosis improved during the follow-up.

Fig. 2.
figure 2

Percentage of CVID patients with acute pneumonia (panel A) and acute otitis (panel B). Black bars: Patients who had acute pneumonias and otitis before the diagnosis and during the follow-up. White bars: Patients with acute pneumonias and otitis during the follow-up only. Grey bars: Patients who had acute pneumonia and acute otitis only before the time of diagnosis.

Full size image
Fig. 3.
figure 3

Percentage of CVID with chronic lung disease (panel A) and chronic sinusitis (panel B). Black bars: Patients who had CLD and chronic sinusitis already at the time of diagnosis and during the follow-up. White bars: Patients who developed CLD and chronic sinusitis during the follow-up. Grey bars: Patients who had CLD and chronic sinusitis at the time of diagnosis only.

Full size image

Chronic Lung Disease.

According to the Italian multicenter study, all patients aged >18 years underwent lung CT scan at the time of diagnosis and subsequently every 4 years. A high prevalence of patients with CLD was present at the time of diagnosis (34.2%). This percentage subsequently increased up to 46.4% during follow-up (Fig. 3, panel A). Fig. 4 shows the age distribution of patients affected by CLD at the time of diagnosis and during follow-up: the prevalence is higher than 50% for almost all adult age groups, but is also surprisingly high, 30–40%, in younger age groups. Chest CT scanning demonstrated the presence of bronchiectasis in 56 patients at the time of diagnosis and in 65 patients at the end of the last follow-up. A unique scoring system to convert the CT image to numeric data has been applied in order to use CT scanning to determine the progression of lung disease in CVID patients (see additional material).

Table II. Prevalence of Autoimmune Diseases at Diagnosis and During Follow-Up in 224 CVID Patients
Full size table
Fig. 4.
figure 4

Prevalence of patients with CLD at diagnosis (black bars) and during follow-up (white bars). All CVID patients enrolled in the study at the time of diagnosis were followed up and grouped in age intervals, as indicated.

Full size image

Autoimmune Diseases.

Table II shows the list of manifestations of autoimmunity observed at diagnosis and at follow-up. Thirty-nine (20 female and 19 males) (17.4%) had autoimmune manifestations prior to CVID diagnosis; 14 of them (35.9%) were younger than 14 years. During the follow-up, 58 patients (29 females and 29 males) (25.9%) had autoimmune diseases; 21 of them (36.2%) were younger than 14 years of age (Fig. 5). Thus, differently from what was expected, there was no difference in the prevalence of autoimmune diseases between female and male patients and between adult and young patients. At the time of diagnosis, in 2.3% of patients, autoimmune diseases were the only clinical manifestation, in 11.1%, autoimmune diseases were associated with recurrent infections, and, in 4% of patients, recurrent infections, autoimmune manifestations, and gastrointestinal diseases were all present. When every type of autoimmune manifestation was analyzed, we found that the prevalence of autoimmune cytopenias decreased during follow-up, due to the treatment with corticosteroids and/or after splenectomy. The observed increased prevalence of other autoimmune manifestations (celiac disease, thyroiditis, vitiligo) during follow-up was probably due to the availability of better diagnostic procedures in patients periodically followed up.

Fig. 5.
figure 5

Prevalence of patients with autoimmune manifestations at diagnosis (black bars) and during follow-up (white bars). All CVID patients enrolled in the study at the time of diagnosis were followed-up and grouped in age intervals, as indicated.

Full size image

Gastrointestinal Diseases.

According to our multicenter study, all patients older than 18 years of age underwent esophagogastroduodenoscopy every 2 years. Signs of chronic gastritis were detectable in 10.2% of patients at diagnosis and in 28.1% at follow-up. Intestinal metaplasia of the gastric mucosa was a common finding. According to the histological and bacteriological findings, patients were treated with antiacid drugs and, in the presence of Helicobacter pylori (H.p.) colonization with eradicating treatments. In the majority of patients, H.p. infection recurred after months or years, requiring new cycles of therapy. Chronic diarrhoea was observed in 14% of patients at diagnosis and in 22.4% of patients during follow-up resulting in a significant malabsorption in 6.2% of patients at the time of diagnosis and in 8.1% at the time of the last follow-up. The percentage of affected patients was significantly higher (p < 0.005 at diagnosis and p < 0.001 at follow-up) in the older age groups (Fig. 6, panels A and B), while there was no difference in the prevalence between males and females. Specific infectious agents have been sought for patients with acute and chronic diarrhoea, but a specific organism (mainly G.lamblia) was documented only occasionally… About 8% of patients had nodular lymphoid hyperplasia and approximately 3% had intestinal granulomatosis. In 3% of patients, gastrointestinal disease was the only clinical manifestation at the time of diagnosis.

Fig. 6.
figure 6

Prevalence of patients with chronic diarrhoea (panel A) and chronic gastritis (panel B) at diagnosis (black bars) and during follow-up (white bars). All CVID patients enrolled in the study at the time of diagnosis were followed up and grouped in age intervals, as indicated.

Full size image

Splenomegaly.

According to our multicenter study, all patients underwent abdominal ultrasound scanning every year. At the time of diagnosis, a spleen enlargement (longitudinal diameter more than 11 cm) was detectable in 17.3% of patients. At the end of the last follow-up, the prevalence of patients with splenomegaly increased to 26.4%. The longitudinal analysis of patient's individual data demonstrated that splenomegaly was not a constant finding: in 9.8% of patients, splenomegaly was present at the time of diagnosis and during follow-up; in 16.5%, splenomegaly was firstly detected during the follow-up, but, in 5.3%, splenomegaly was detectable only at the time of diagnosis and the spleen size returned to normal size during the follow-up. Six patients have been splenectomized during follow-up because of hypersplenism. In all but one, splenectomy successfully cured the cytopenia. Splenomegaly was always associated with a liver enlargement (longitudinal diameter of left lobe more than 12 cm) and/or abdominal lymphoadenopahy. In 5.6% of patients, splenomegaly was associated with leukocytopenia (<4000/mm3) and thrombocytopenia (<30000/mm3), in 3% with thrombocytopenia only, and in 2% with leukocytopenia only. We did not find any difference in the prevalence of splenomegaly between males and females and between adults and children at the time of diagnosis. A lower prevalence of splenomegaly was found in children in comparison to the adult groups during follow-up (p < 0.005).

Lymphoma and Cancer.

Cancers developed in 14 patients. Table III shows the list of the cancers, the SPR, and the 95% CI. The increased risk to develop cancers was due to a significant increased risk to develop non-Hodgkin's lymphoma (NHL) (four patients, SPR 18.61) and stomach cancer (two patients, SPR 16.22). All lymphomas were B-cell type. All patients received chemotherapy and anti-CD20 treatment. One patient underwent autologous bone marrow transplantation.

Table III. Tumours Observed in CVID Patients with SPR and CI 95%
Full size table

Treatment with Immunoglobulins

At the time of diagnosis, the mean baseline serum IgG was 258.12 mg/dL, the mean level of IgA was 27.56, and the mean level of IgM was 40.06 mg/dL. After 1 year of IVIG treatment, the mean baseline serum IgG was 579.49, the mean baseline serum IgA was 17.2 mg/dL, and the mean level of IgM was 29.76. No differences were found in the Ig serum levels between males and females. All patients were receiving IVIG from five commercial sources (branch A: 63.5%, branch B: 16.3%, branch C: 13.2%, branch D: 4.5%, branch E: 2.5%) with dosage and timing established according to the Italian recommendation (18): all patients received a monthly dosage of 400 mg/kg weight of IVIG at intervals ranging between 2 and 3 weeks. The majority of patients (79%) started IVIG replacement therapy at the time of diagnosis or within 1 year from diagnosis (13%). Four percent of patients started IVIG therapy between 2 and 5 years from diagnosis and 4% between 5 and 10 years from diagnosis. The reasons for the delay of IVIG treatment were the absence or paucity of symptoms (10 patients) and the worry for the risk of viral transmission (six patients). In the majority of patients (58%), the IVIG monthly dosage remained stable during follow-up. In 31% of patients, the IVIG dosage has been increased because of weight increase (18%) or because of a poor control of lower respiratory tract infections (13%). In 11% of patients, the monthly IVIG dosage has been reduced. The trough IgG levels do not correlate with development of chronic diseases (chronic sinusitis, chronic lung disease, chronic gastritis, and chronic diarrhoea in our series (manuscript in preparation)).

Antibiotic Prophylaxis.

11.6% of patients were treated with a daily antibiotic prophylaxis with trimetropin-sulphabactan (8%) or clarytromicin (3.6%). Antibiotics were used in all patients with acute exacerbations of CLD characterized by worsening of dyspnoea, increased sputum volume, and purulence with or without a demonstrable bacterial infection

Survival.

The median follow-up for CVID patients was 11.5 years. During the study, 13 patients died, from 3 to 12 years after diagnosis, at ages ranging from 3 to70 years (median age 42) (Fig. 7). The major single case of death was cancer (six patients). Another important cause of death was CLD (four patients). Two patients died of liver failure, and one of accident.

Fig. 7.
figure 7

Kaplan–Meier survival curves after diagnosis of CVID in the entire cohort.

Full size image

DISCUSSION

This multicenter prospective study on 224 Italian patients with CVID provides an updated view of illnesses occurring at the time of clinical onset and over a long period of follow-up (11.5 years). Moreover, we provide data on the effect of long-term immunoglobulin replacement therapy. Similarly to data reported in 1999, in a large study of 248 patients with CVID (8), a wide spectrum of associated clinical conditions, including respiratory tract infections, autoimmune diseases, gastrointestinal manifestations, and tumors were found. In Italian CVID patients, the male:female ratio was 1:1. The average age at diagnosis was 26.6 years, lower than that reported in previous studies; the mean time of diagnostic delay was 8.9 years, longer than that previously reported (7, 8). The delayed recognition of the antibody defect, possibly due the variability of the first clinical manifestations of CVID, was one of the possible reasons leading to the high prevalence of CVID patients with CLD (30%) and to a high prevalence of patients with chronic sinusitis (36%) detected at the time of diagnosis. Disappontingly, despite IVIG treatment, the number of CVID patients with CLD and chronic sinusitis further increased during follow-up for almost all age groups, including children. This figure was similar to that we have reported in XLA patients (19), where the overall probability of developing CLD reached approximately 80% after 17 years of follow-up. Abnormal CT scan results proved to be a predictor for lung disease progression and for the development of bronchiectasis. As already demonstrated in patients with cystic fibrosis, dyspnea and sputum production are conditioning factors of increased morbidity (20). Accumulated mucus in the airways is the prominent feature of bronchiectasis leading to airway obstruction, bacterial colonization, and recurrent infections.

Only a minority of our patients were under antibiotics prophylaxis. According to the Italian network recommendations (19), we treated all patients with exacerbations of bronchiectasis with antibiotics, with the choice of agents depending on the likely causative pathogens.

Our study demonstrated a high prevalence of chronic gastrointestinal diseases, chronic gastritis, and chronic diarrhoea emerging mainly in older age groups, and leading to malabsorption in a significant number of patients. Different explanations may be offered to justify the high prevalence of these chronic conditions in patients with humoral immunodeficiencies under appropriate immunoglobulin replacement therapy. IVIG may, in fact, substitute IgG only, while IgA and IgM, that are the major secretory antibodies at mucosal surfaces, cannot be replaced. Therefore, the progression of chronic respiratory and chronic gastrointestinal diseases may still occur in hypogammaglobulinemic patients. On the contrary, the percentage of CVID patients affected by acute respiratory infections, pneumonia, and otitis significantly decreased during the follow-up, when patients were receiving IVIG therapy. Moreover, invasive bacterial infections, meningitis, and sepsis, were rarely observed in our patients and only in those patients who refused IVIG treatment. Surprisingly, the long-term clinical follow-up of CVID patients has led to the observation that, despite severe hypogammaglobulinemia, some CVID patients do not suffer from recurrent respiratory tract infections and a minority of them were almost asymptomatic. A possible explanation for these findings has been recently clarified (21). A constant level of antibodies is indispensable to ensure protection against infections. Serum immunoglobulins include low- and high-affinity antibodies of different isotypes, either produced without previous immunization (natural antibodies) or resulting from antigenic challenge (adaptive antibodies) (22). Natural antibodies are mostly of IgM isotype and represent a promiscuous repertoire: thanks to the combination of low-affinity and high-avidity, IgM are an efficient first-line defence against all invading pathogens (23). Adaptive antibodies, instead, are the final product of the complex response to antigen occurring in the germinal centers through somatic mutations, affinity selection, and class-switch recombination (22). They specifically and exclusively recognize their target pathogen and have the function to clear the infection and to prevent its recurrence. The clinical variability observed in CVID patients, including the apparent paradox of patients with severe hypogammaglobulinemia, without recurrent infections, might be explained by a different residual B-cell function resulting in the differentiation of a B-cell subset, the so called “IgM memory B cells,” and the production of detectable levels of antipolysaccharide antibodies (21). The innate immunity in CVID patients with a paucity of symptoms is preserved: they may produce low but detectable levels of IgM, while the defective adaptive immunity may be compensated by the IgG substitutive treatment (21). However, in CVID patients with CLD, this first line of defence is lacking. They do not produce any residual level of IgM, and IVIG cannot compensate this defect (21). Because it has been suggested that trough IgG serum levels >800 mg/dL may lower the incidence of serious bacterial infections (17–19), we increased the monthly IVIG dosage in about 13% of patients. However, the trough IgG levels at follow-up do not correlate with development of chronic diseases in our series (manuscript in preparation). It is possible that patients who continue to have respiratory infections and who develop CLD may take advantage of a more aggressive therapeutic strategy, such as prophylaxis of infectious episodes with antibiotics (24), and continuous cycles of respiratory rehabilitation. Prevention of chronic gastrointestinal diseases and malabsorption might also be achieved by a closer medical and instrumental monitoring.

Beside recurrent infections, other clinical conditions were associated to CVID. Autoimmune manifestations were observed prior to CVID diagnosis in 14% of patients. In 2% of patients, autoimmune diseases were the only clinical manifestation at diagnosis. Contrary to what was expected, there was no difference in the prevalence of autoimmune diseases between female and male and between adult and young patients, suggesting that autoimmune diseases are a part of the CVID clinical spectrum of illnesses and not a simple association. While the prevalence of autoimmune cytopenias decreased during follow-up, we observed an increased prevalence (29%) during follow-up, mainly due to the increased number of patients diagnosed with celiac disease, misdiagnosed for a long time. Only in a minority of patients, the autoimmune manifestations account for the condition of spleen enlargement observed in a high percentage of patients. Splenomegaly, already described in other studies, has been associated with B and T-cell abnormalities in a subgroup of CVID patients (25–28). In order to prove the association between splenomegaly and CVID in a particular subset of patients, we asked the question of whether the spleen enlargement was a constant finding. The longitudinal ultrasound analysis of patient's individual data demonstrated that in 26% of patients the spleen size constantly increased during follow-up, while in about 5% splenomegaly was detectable only at the time of diagnosis and then returned within the normal parameters. In the majority of patients, splenomegaly was associated with a liver enlargement and/or abdominal lymphoadenopahy. In about 6% of patients with splenomegaly, the histological staining of lymph nodes, liver, and spleen demonstrated a granulomatous reaction, as described (29–31). We did not find any difference in the prevalence of splenomegaly between males and females. A multivariate analysis did not demonstrate, at the moment of the last follow-up, any association between splenomegaly and other clinical or laboratory abnormality.

We found 14 patients with tumours with a significant increase in the SPR of 3.22: four were NHL (SPR: 18.6) and two stomach cancers (SPR: 16.2). The predictive relative risk for tumours in CVID patients described in previous studies was variable. The Immunodeficiency Cancer Registry collected more that 500 cases of tumours in patients affected by Primary Immunodeficiencies, with 48.6% of NHL (32). In the 1999 study on CVID patients (8), about 8% developed NHL and about 1% stomach cancers. In a British study on 220 CVID patients (33), 14 developed tumours with a cumulative relative risk of 5: seven were stomach cancers and three NHL. In a recent Danish–Swedish study on 176 CVID patients (34), the relative risk was 1.8, mainly due to an increased relative risk of NHL and stomach cancers.

The mortality we observed in our CVID patients is much lower, about 6% over a median follow-up of11 years, than that previously reported in the 1999 study (between 23 and 27% over a median follow-up of 7 years) (8). This difference may reflect the impact of IVIG treatment that became the standard treatment after the beginning of '80s years. The parameter most strongly associated with survival in the series reported in 1999 was the percentage of peripheral B cells (8). It is possible to hypothesize that, after almost a decade, both IVIG treatment and better diagnostic and therapeutic strategies, had a great impact on CVID mortality, asalready reported in patients enrolled in a parallel prospective study on Italian XLA patients (19). As discussed earlier, B-cell abnormalities may still be the parameter most strongly associated with the development of chronic diseases and prognosis in CVID patients (21, 25–28, 35). Our data stressed the need to develop international guidelines for the prevention and therapy of chronic diseases in patients with humoral immunodeficiencies.