Health status and infections in patients with symptomatic primary and secondary immunoglobulin G (IgG) deficiencies receiving intravenous IgG substitution CURRENT STATUS: UNDER REVIEW

Background: The effects of infections of patients from immunoglobulin G (IgG) deficiencies evaluated in a prospective analysis. Methods: Patients with symptomatic primary or secondary IgG deficiencies were interviewed prior to the first IgG infusion (t0) and over the course of their treatment. The respondents rated their current health using a 100-point scale (EQ-5D-5L), ranging from 0 ('worst imaginable health') to 100 ('best imaginable health'). The patients also provided information on the frequency of infections and infections requiring antibiotics in the past 8 weeks. A control group (CG) without oncologic diseases answered the questions on an one-off basis. Results: 106 patients with a median age of 65 years (21-85 years) were investigated. The median serum IgG-concentration changed from 500 mg/dl (t0) to 772 mg/dl (t6). The mean number of infections and of infections requiring antibiotics decreased during IgG substitution significantly. The EQ-5D-5L health status improved from 57 (t0) to 68 (t6), compared to 73 in the CG. Conclusion: During the course of IgG substitution patients reported fewer and less severe infections. Their health status improved but still was inferior to the healthy CG.


Introduction
The administration of human polyvalent immunoglobulin G (IgG) is an effective way of preventing infections. Substitution is primarily indicated for primary immunodeficiency disorders (PID) with a limited production of antibodies, which is frequently the result of a genetic defect. It is also indicated for patients with hypogammaglobulinemia as well, as those suffering from recurrent bacterial infections due to a chronic lymphoproliferative disease such as chronic lymphocytic leukemia (CLL) and multiple myeloma (MM). Another field of application is hypogammaglobulinemia in stem cell transplants irrespective of preceding infections.
The potentially life-threatening risk of infection through bacteria such as streptococcus and haemophilus influenzae, which are normally kept under control through antibody response, is clinically important particularly in the case of patients with immunodeficiencies. If the patients are found to suffer from such infections, therapy with antibiotics or IgG is indicated in order to prevent organ damage and death (1). Randomized controlled studies have shown that intravenous IgG (IVIG) substitution lowers the infection rate for CLL significantly (2)(3)(4)(5). A literature analysis that looked at randomized controlled studies showed that the risk of patients with lymphoproliferative diseases such as CLL and MM developing interstitial pneumonia was reduced significantly when they were treated with polyvalent IgG and that clinically and microbiologically documented infections had decreased.
There was however no proof of IVIG substitution resulting in lower mortality rates (6).
In medical practice, IVIG therapy with IgG substitution is frequently indicated for oncology patients with symptomatic secondary hypogammaglobulinemia.
Our report covers a prospective health status analysis of patients with symptomatic IgG deficiencies who were given IVIG at oncology group practices in Germany. We also looked at the benefits of IgG replacement therapy in order to reduce the number and severity of infections. A comparison of patients to a control group (CG) and a sub-analysis of patients with primary and secondary immune defects completed the analysis.

Methods
Patients with symptomatic primary or secondary IgG deficiency who were about to start IVIG therapy were included. Overall, 12 sites took part in this multi-centre study in Germany. Patients received IgG products supplied by different manufacturers.
Interviews with the patients took place at the start of the treatment and were repeated at 8-weekly intervals during up to six measuring times. Treatment data and assessments of the oncologists in charge of treatment were linked with the data gained from the interviews. In order to be better able to assess the results of the patient survey, a one-off survey was conducted in an age-adjusted healthy CG with a similar sex distribution. 'Healthy' in that context meant that the respondents had no malignant or immunodeficiency disease.
Patients and CG assessed their current state of health based on a validated 100-point scale (EQ-5D-5L) ranging from 0 ('worst imaginable health') to 100 ('best imaginable health'). In addition, some questions were processed using a non-validated tool with scale end points 1 being 'very good' to 5 being 'not good at all / bad' and 1 being 'never' to 5 being 'always', respectively. 47 patients dropped out of the project over the course of the observation period. In most cases, an IgG substitution was no longer necessary in these patients, but in some cases there was no longer a willingness to participate.
The data of these patients were analyzed for all measurement times at which values were available.
No patient was excluded from the analysis due to incomplete data.
Data was analyzed with the Statistical Package for the Social Sciences (SPSS) 19. Frequencies, medians, means and standard deviations were calculated to describe the data. Data was analyzed for all patients and the two subgroups PID and SID. Patients suffering from multiple myeloma were excluded from analyzes of serum IgG level due to the disease-related production of monoclonal Ig (Mprotein). Furthermore IgA and IgM were not assessed for patients suffering from multiple myeloma and Waldenström macroglobulinemia respectively.
T-tests for two independent samples were used to check the mean differences in age between patient and control group and between the patient subgroups for statistical significance. Three repeated measures ANOVA (analyzes of variance) were conducted to analyze the development of infections, infections requiring antibiotics and health status according to the EQ-5D-5L scale in the course of the treatment (t 0 -t 6 ). Single missing values did not exclude patients from these analyses. ANOVA were not analyzed regarding normal distribution since simulation studies have shown that the ANOVA with repeated measurement is relatively robust against violations of the normal distribution assumption (7). The Greenhouse-Geisser adjustment was used to correct if violations of sphericity occurred.
Bonferroni-adjusted post-hoc analyses were used to check the differences in mean values in the course of the treatment (t 0 -t 6 ).

Results
The majority of patients (73%) was given IgG substitution due to a secondary antibody deficiency, 27% due to a PID. Patients with PID were statistically significantly younger than patients with SID (mean age 52 years vs. 66 years; p < .001). The sex distribution also differed. 72% of the patients with PID were female compared to 36% of the patients with SID.
Patients with secondary antibody deficiency received IgG substitution due to chronic lymphocytic leukemia (35%), multiple myeloma (22%), follicular lymphoma (6%), other non-Hodgkin lymphomas (30%) or other malignant diseases (6%). 58% of the patients with secondary antibody deficiency were not given immunosuppressive treatment. The most frequently used immunosuppressive treatments provided to the remaining patients were rituximab (19%), corticosteroids (18%) as well as immunomodulatory drugs (lenalidomide and pomalidomide with 8%). In the last 2 months, 95% of the patients had suffered from infections and 81% had suffered from infections requiring the use of antibiotics.
The median immunoglobulin values in the serum at time t 0 amounted to 500 mg/dl for IgG (n = 88, without MM patients), 62 mg/dl for IgA (n = 87, without MM patients) and 26.5 mg/dl for IgM (n = 98, without patients suffering from Waldenström macroglobulinemia). Patients with PID had a median serum IgG of 600 mg/dl at t 0 compared to 460 mg/dl in patients with SID (without MM patients).
The scheduled treatment intervals were 4 weeks in 80%, 3 weeks in 8% and 2 weeks in 6% of the cases. The mean IVIG dose infused per 4 weeks was 21 g (10 g − 70 g). Patients with PID had a mean dose of 19 g compared to 21 g in patients with SID.
Changes in the course of the IgG substitution

IgG levels
Under substitution, the median IgG value rose from 500 mg/dl (t 0 ) to 772 mg/dl (t 6 ) and was therefore at the lower end of the normal range. Patients with PID had an IgG of 600 mg/dl at t 0 and of 815 mg/dl at t 6 . Patients suffering from SID (without MM patients) had a median IgG value of 460 mg/dl prior to substitution and an IgG value of 751 mg/dl at t 6 . Figure

Infections and infections requiring antibioticssubgroups
Patients with PID reported more infections than patients with SID. In both groups a substantial decrease in the number of infections could be observed after initiation of IgG substitution.
Patients with SID had a mean number of 1.6 infections before substitution. The mean values developed as follows: 1.0 (t 1 ), 0.8 (t 2 ), 0.6 (t 3 ), 0.6 (t 4 ), 0.4 (t 5 ) and 0.6 (t 6 ).   According to actual guidelines of the European Medical Agency (EMA) an IgG substitution therapy is not indicated until the patient shows a significant antibody deficiency and/or shows proven specific antibody failure and suffers, in addition, from recurrent severe infections refractory to antibiotic therapy. Furthermore, antibiotic therapy options must have been exhausted or is not indicated (1).
In medical practice, we often see that most patients have more than 2 infections per year needing antibiotic therapy. Based on our experience, a significant percentage of symptomatic patients with indolent lymphomas benefits from a targeted serum-level adjusted IgG substitution therapy. A clear reduction of infections requiring the use of antibiotics and an improvement of the patients' quality of life serve as proof (9).
The normal therapy schedule consists of four-weekly administrations of 30 g of IVIG. The aim is to achieve serum levels of 700 to 1,000 mg/dl. Based on our experience, higher IgG levels are not necessary as they would not offer the patients a measurable clinical benefit. Previous analyses, too, have shown that IgG levels between 700 and 1,000 mg/dl are sufficient to reduce the frequency of severe infections. In this analysis, a median serum level of 772 mg/dl, i.e. a value at the lower end of the normal range was achieved, which may be due to the fact that more than half of the patients had only been given 10 g (38%) and/or 20 g (17%) of IgG per treatment cycle. The results presented in this context reflect the everyday experience at medical practices in Germany.
When comparing our data with that of other analyses, the question remains as to whether a dose of 30 g (or 0.4 g/kg body weight) for patients with repeated infections could possibly have led to a better result (2,3,8,9). Recently a cohort study of 8,633 patients from a tertiary referral center receiving rituximab has been presented. Rituximab led to an increase in severe infections. 85% had no measurement of immunoglobulin levels before rituximab therapy. Only 4.5% received immunoglobulin replacement therapy. Higher cumulative replacement dose was associated with a reduced risk of severe infections (10).

Improvement of health status
Based on the EQ-5D-5L, the patients' health status improved on average from 57 (t 0 ) to 68 (t 6 ). 64% of the doctors in charge of treatment claimed that their patients had benefitted 'greatly' from the substitution treatment whilst 30% said that they had benefitted 'a little' and according to 6% patients had benefitted 'not at all'. At time t 6 65% of the patients had continued with the IgG substitution while 35% had stopped, due to a variety of reasons (antibody deficiency no longer required treatment, patient rejected further IgG substitution, side effects, patient died).
As far as we know there has, up to now, been very little data about the health status of patients receiving IgG substitution. This prospective analysis confirms the results of a prospective survey published in 2015 concerning patients with indolent non-Hodgkin lymphomas and symptomatic antibody deficiency (9). It also showed that the rise of IgG serum levels led to fewer infections requiring the use of antibiotics (9).
When comparing the health status between patients and CG we observed an improvement in the patients from 57 (t 0 ) to 68 (t 6 ) during IVIG-treatment, but it was still inferior to the healthy CG who scored 73 on the EQ-5D-5L. This shows that, despite IVIG-treatment, patients' health status remains inferior to persons without a malignant disease, which is multifactorial and probably partly caused by chronic fatigue. Large cohort studies have shown that patients with CVID and long term survivors of malignant lymphoma suffer from increased chronic fatigue (11,12).

Methodological considerations
The present study is a prospective multi-center patient survey. In addition, medical data, such as the IgG level in the serum, were transferred from the treatment files and the treating oncologists assessed the success of treatment with the last IgG administration. The data were linked and processed and are probably representative of the treatment reality of patients in hematological and oncological group practices in Germany. When carrying out the project, however, methodological compromises had to be made. Methodological limitations result from the rather small population of 106 patients. There was also no a priori calculation of the sample size. From a practical point of view, however, it must be noted that these patients are rare and that they had to be interviewed before the first IgG administration. A comparison of the patient subgroups with primary vs. secondary immunodeficiencies was therefore only possible at a descriptive level. The patient group may not have been homogeneous enough, although the results found seem to apply to both subgroups. A possible problem that the patients form their own control group by repeating the measurement applies. Nevertheless, we wanted to use a "gold standard" from healthy controls in order to better assess our results. For practical and financial reasons, this group could only be interviewed once. A comparison with the results of this control group is therefore only possible to a limited extent. When interpreting the data of the CG, it must also be taken into account that it was interviewed in August and therefore may have reported less infections and infections requiring antibiotics.
Perhaps the biggest methodological problem is the comparatively high drop-out rate of 35%, i.e. 35% of all patients did not provide data at all times. Most of these patients could not be interviewed because, from a clinical point of view, they no longer needed any further substitution therapy. This creates a negative selection for our project, since patients who have responded particularly well to the substitution are the first to drop out. Overall, the effects found seem to be so robust that this drop-out can be neglected.

Declarations
Ethics approval This study was approved by the ethics committee of Rhineland-Palatinate, Germany.

Consent for publicationl
Not applicable Availability of data and materials The datasets used and analyzed during the current study are available from the corresponding author on reasonable request. Health status (mean values) in the course of the replacement therapy (t0 -t6)