British Society for Immunology/United Kingdom Primary Immunodeficiency Network consensus statement on managing non-infectious complications of common variable immunodeficiency disorders.

Common variable immunodeficiency (CVID) represents a heterogeneous group of rare disorders. There is considerable morbidity and mortality as a result of non-infectious complications, and this presents clinicians with management challenges. Clinical guidelines to support the management of CVID are urgently required. The UK Primary Immunodeficiency Network and the British Society for Immunology funded a joint project to address this. A modified Delphi Survey was conducted for the assessment, diagnosis and treatment of the non-infectious blood, respiratory, gut and liver complications of CVID. A steering group of 10 consultant immunologists and one nurse specialist developed and reviewed the survey statements and agreed the final recommendations. In total, 22 recommendations and three areas for research were developed.

Introduction

Common variable immunodeficiency (CVID) represents a heterogeneous group of rare disorders, with prevalence ranging from 1 in 10 000 to 1 in 50 000 in different populations 1, 2. CVID is characterized by hypogammaglobulinaemia, defective specific antibody production and increased susceptibility to recurrent bacterial infections 2, 3, 4. In addition to the increased risk of recurrent infections, patients with CVID are also at greater risk of autoimmune disorders and cancer 1, 5, 6. Immunoglobulin replacement is the main treatment for the prevention of recurrent infections; however, there is little evidence regarding treatment options for the non‐infective complications of CVID 3, 7, 8.

There is considerable morbidity and mortality as a result of non‐infectious complications, and this presents clinicians with management challenges. Further clinical guidelines to support the management of CVID are urgently required. The UK Primary Immunodeficiency Network and the British Society for Immunology funded a joint project, facilitated by the National Guidelines Centre (NGC), to address this. A comprehensive search of the literature undertaken by the NGC revealed little evidence or clinical guidance on the management of other non‐infectious complications of CVID. The lack of evidence reflects the challenges posed in CVID and other rare diseases, where meaningful randomized controlled studies are extremely difficult to conduct.

The published evidence available to date did not support the development of a formal evidence‐based guideline in this area, so a Delphi consensus process (anonymous, multi‐round, consensus building technique) was adopted to address the assessment, diagnosis and treatment of the non‐infectious blood, respiratory, gut and liver complications of CVID. The Delphi method was developed in the 1950s, and has been used successfully for generating, analysing and synthesizing an expert view to reach a group consensus position. The Delphi process relies upon a group of experts (in this case, all consultant immunologists and specialist immunology nurses in the United Kingdom were invited to participate) responding to and providing critical feedback on two or more consecutive surveys (in this case, a set of statements regarding the management of non‐infectious complications of CVID). The initial statements were developed by a steering group of 10 consultant immunologists and a nurse specialist. Following the Delphi principles, the initial recommendations that had not met consensus after the first round of the survey were modified according to the expert feedback and recirculated to the expert group. The recommendations that achieved consensus group are described in the Results section of this paper. The group were aware of the consensus statement on granulomatous–lymphocytic interstitial lung disease (GLILD) 9 and therefore did not repeat this work.

Methods

Delphi methods

Development of consensus statements

The steering group met and identified the non‐infectious complications and the areas of management to be included within the survey. The steering group agreed to focus on non‐infectious blood, respiratory, gastrointestinal (GI) and liver complications as the most commonly encountered complications where there is little evidence to guide practice. The steering group formulated and validated the consensus statements at each round for the survey. The statements were revised if further clarification was required and sent back to the expert group until no further clarification was needed.

Responses were graded on a four‐point Likert scale: ‘strongly agree’, ‘agree’, ‘disagree’ and ‘strongly disagree’, with an option of ‘I don’t have the expertise’. A threshold of 70% to agree (strongly agree or agree) or disagree (strongly disagree or disagree) was used to evaluate the responses. This threshold is routinely used in Delphi processes as a recognized measure of consensus, and statements reaching this threshold were noted to be agreed and not included in subsequent rounds 10. Free text boxes were included in the survey after each section to give respondents the opportunity to comment on or clarify the statements or responses. These comments were taken into account when revising the statements that did not meet consensus. There is no complete agreement about the termination of a Delphi process, and when no clear indication was given on how to amend a statement from the expert group this was removed from the survey.

The expert group for the survey was identified from the Royal College of Physicians (RCP) 2016 consultant census of immunologists. The aim was to target all the consultants practising in clinical immunology adult services in the United Kingdom. The census identified 85 consultants, in addition to the members of the steering group. The steering group did not participate in the Delphi Survey. The RCP sent out invitation e‐mails and two reminders for each round of the Delphi survey. E‐mails were sent to each identified consultant outlining the aim of the survey, giving instructions and a link to the survey. Nurse specialists were also invited to take part. The nurse specialists were identified from the Immunology and Allergy Nurses group, 75 of whom identify as immunology specialist nurses. The survey questionnaire was sent to all members of the expert group via SurveyGizmo software. Responses remained anonymous. The whole process was conducted online, with an 8‐week cycle time between rounds. This allowed for the questionnaire invitation to be sent, for two reminders to be sent to participants for completion and for the analysis to be conducted prior to the next sequential round.

The steering committee formulated recommendations based on the consensus statements. The steering committee reviewed all free text comments from the responders and these are discussed in each section of the recommendations.

Results

Demographics

The steering group that developed the questionnaire consisted of a facilitator from the NGC, 10 consultant immunologists and one nurse specialist.

Thirty‐nine participants from 21 immunodeficiency services (of a total of 34 adult UK immunodeficiency services) took part in the first round of the survey; 35 of 85 (41%) consultants and four of 75 (5%) nurse specialists. The first‐round survey was completed by 33 of the 39 participants, while six participants only partially completed the survey. More than 70% of the participants had worked in immunology as a speciality for more than 10 years, and each had direct involvement with more than 50 patients with CVID. Twenty‐two participants, all of them consultants, from 16 immunodeficiency centres completed the second round of the survey. More than 80% of the participants had worked in immunology as a speciality for more than 10 years and 50% have direct involvement with more than 50 patients with CVID.

Of the 63 statements in the first round, 57 achieved consensus. For details of the statements that did not reach consensus see Appendix 1. Of the remaining six statements, two still did not achieve consensus in the second round. The steering group made the decision not to make recommendations based on the two statements that did not reach consensus. The statements that achieved consensus were reviewed by groups representing patients with CVID, and their comments have been incorporated into the discussions below.

Summary of recommendations

The recommendations should be considered in the context of the individual needs of the adult with CVID.

Baseline assessment and monitoring for non‐infectious complications

Although not all people with CVID develop complications, close attention to monitoring is required to identify any complications at an early stage. It is not yet possible to confidently predict the subsets of people at high risk.

Recommendation 1. Adults with CVID should:

be weighed at least 6‐monthly

have a clinical examination of their lymph nodes at least annually to monitor for the development of persistent lymphadenopathy

have a clinical examination of their abdomen at least annually to monitor for the development of hepatosplenomegaly

have a full blood count taken 6‐monthly to monitor for the development of autoimmune cytopaenias

have blood tests for B12, folate and ferritin in the presence of anaemia, malabsorption or neuropathy

have liver function tests (LFTs) assessed 6‐monthly to monitor for the development of hepatitis

have full pulmonary function tests including transfer factor at least every 3 years to monitor for the development of interstitial lung disease

have a baseline high resolution CT (HRCT) chest to exclude bronchiectasis and lymphadenopathy/other pathology. If a HRCT chest has been performed in the last 12 months, this can be used as the baseline assessment:

a baseline HRCT chest should include the upper abdominal viscera to assess lymphadenopathy, spleen size and hepatobiliary architecture unless an abdominal ultrasound has been performed in the last 12 months to assess the liver and spleen.

Eighteen per cent (n = 7) of respondents suggested that full pulmonary function testing was performed annually for all patients in their practice. Recommendation 8 increases the frequency of pulmonary function testing to annually in CVID patients who are found to have splenomegaly.

Multi‐professional care

Recommendation 2. Adults with non‐infectious complications of CVID should be managed as part of a multi‐professional team that includes an appropriate organ‐based specialist with an interest in primary immunodeficiency.

Haematological complications

Recommendation 3. Where an autoimmune cytopenia is suspected:

a clinical assessment for other secondary causes should be undertaken (for example, drugs, hypersplenism or lymphoproliferative disease)

a blood film should be a first‐line investigation to help identify the underlying cause

antibody testing (for example, for anti‐granulocyte, anti‐platelet) should not be performed routinely

Recommendation 4. Immunology centres should agree monitoring, action or referral thresholds for adults with CVID and features of haematological complications with the local haematology team.

Recommendation 5. Adults with autoimmune haemolytic anaemia should be treated according to the British Society for Haematology Guidelines on the management of drug‐induced immune and secondary autoimmune haemolytic anaemia 11.

Respiratory complications

There is an overlap between infectious and non‐infectious respiratory complications of CVID. For this reason, the Steering Group decided to include statements regarding identification and management of bronchiectasis. The management of upper respiratory tract infections is not addressed in this guideline.

Recommendation 6. Adults with CVID and ongoing frequent respiratory tract infections should have a HRCT chest at least every 5 years to monitor for the development of bronchiectasis.

Recommendation 7. Adults with CVID and bronchiectasis and ongoing frequent respiratory tract infections should have a HRCT chest at least every 5 years to monitor the progression of bronchiectasis.

Recommendation 8. Adults with CVID and splenomegaly should have annual pulmonary function tests (PFT) including transfer factor to monitor for the development of interstitial lung disease.

Recommendation 9. Centres should agree monitoring, action or referral thresholds for adults with CVID with features of respiratory complications with the local Respiratory team.

Recommendation 10. Adults with CVID and bronchiectasis should be managed jointly with a consultant chest physician and an immunologist.

In the free text section, all the respondents stressed the importance of using clinical judgement and individual patient assessment to determine the exact frequency of computerized tomography (CT) scanning and pulmonary function tests (PFT). It is acknowledged that in some cases more frequent scanning or PFT may be indicated. The radiation risk associated with CT scanning was highlighted by respondents commenting that some patients with CVID may be particularly radiosensitive, and thus potentially more susceptible to ionizing radiation. The role of PFT in place of, and as an adjunct to, HRCT was emphasized by several respondents in order to minimize radiation exposure. The committee concluded that alternatives such as low‐dose HRCT and magnetic resonance imaging (MRI) may also be considered.

GI complications

Infectious diarrhoea is the most common GI complication in CVID 12, while 9–15% of patients are reported to develop non‐infectious GI complications 8, 13. Diagnosis of non‐infectious GI complications requires that infection is excluded as a cause of symptoms. As a result, the steering group agreed to include statements relating to infection in the Delphi survey, even though infectious complications are not the main focus of these guidelines.

Recommendation 11. Adults with CVID and evidence of malabsorption should have the following investigations:

haematinics

calcium

vitamin D

fat soluble vitamins

Respondents highlighted that malabsorption should be considered in patients with unexplained weight loss.

Adults with CVID and Helicobacter pylori

Recommendation 12. Adults with CVID and H. pylori should be offered eradication in line with National Institute for Health and Care Excellence (NICE) guideline CG184 (Gastro‐oesophageal reflux disease and dyspepsia in adults: investigation and management) 14.

Adults with CVID and diarrhoea

Recommendation 13. In adults with CVID and diarrhoea for more than 2 weeks the following infections should be excluded:

Campylobacter

Clostridium difficile toxin

Giardia

Salmonella

Shigella

Cryptosporidia

Microsporidia

Cytomegalovirus (CMV)

Norovirus

Other intestinal parasites (for example, isospora)

In the free text section, a number of respondents favoured a stepwise approach, with first‐line investigations to exclude Campylobacter, C. difficile toxin, Giardia, Salmonella and Shigella, only proceeding to investigating for viruses, mycobacteria and other parasites if these first‐line tests are negative.

Recommendation 14. Adults with CVID and non‐infectious diarrhoea or malabsorption should be referred to a consultant gastroenterologist and a dietician.

Recommendation 15. Where initial tests for infection are negative, all patients with persistent diarrhoea should be referred for lower GI endoscopy with biopsy for the assessment of:

GLILD – including typical and atypical Crohn’s disease

inflammatory bowel disease, including typical and atypical ulcerative colitis

lymphocytic infiltration to suggest autoimmune enteropathy

lymphoma (T and B cell)

infection [CMV, enteroviruses, Epstein–Barr virus (EBV)]

Recommendation 16. Where initial tests for infection are negative, all patients with persistent diarrhoea should be referred for upper GI endoscopy with biopsy for the assessment of:

coeliac and coeliac‐like enteropathy

lymphocytic infiltration to suggest autoimmune enteropathy

small bowel biopsy and jejunal aspirates to exclude giardia

lymphoma (T and B cells)

infection (CMV, enteroviruses)

nodular lymphoid hyperplasia

Recommendation 17. Centres should agree a protocol with the local gastroenterology team to guide observations and indicate the pathological samples to be taken at endoscopy in adults with CVID.

The consensus was clear that both upper and lower GI endoscopies should be examined to exclude lymphoma, and respondents emphasized the importance of working closely with histopathologists and gastroenterologists.  The respondents recognized the lack of evidence regarding the underlying reasons for the increased risk of malignancy in CVID, but there was clear consensus for the need to exclude malignancy in this clinical context. Both rounds of the survey included statements which included specific methods for identification of viruses and CD4 and CD8 T cells in gastrointestinal biopsies, and these all received high levels of agreement. However, the expert group felt that these were too restrictive, and that the intent of the recommendations was to highlight the need to exclude specific infections and lymphoma without limiting the methodologies used. The statements were condensed to reflect this.

Consensus was not reached regarding examination of stool samples for EBV and AAFB; however, it was recommended that if biopsies are taken EBV should be excluded.

Recommendations for adults with CVID and pernicious anaemia

Recommendation 18. Adults with CVID and pernicious anaemia should have:

an upper GI endoscopy performed within 6 months of diagnosis

monitoring arranged according to local gastroenterology guidance

Recommendations for adults with CVID and coeliac disease

Recommendation 19. Adults with CVID and coeliac disease should have:

an upper GI endoscopy and biopsies performed prior to gluten avoidance

an upper GI endoscopy and biopsies performed after 6 months on a gluten‐free diet to assess response

monitoring arranged according to local gastroenterology guidance

The steering group acknowledge the difficulty in diagnosing coeliac disease in patients with CVID in whom the serological tests are not useful. The existence of non‐gluten sensitive enteropathy histologically similar to coeliac disease is also recognized 15. Comments from the survey highlighted the possibility of using human leucocyte antigen (HLA)‐DQ2/8 testing to identify patients in whom coeliac disease is unlikely.

Hepatobiliary complications

Recommendations for adults with CVID and abnormal liver function tests (LFTs)

Recommendation 20. Adults with CVID and abnormal LFTs should have:

LFTs repeated to avoid unnecessary additional investigations

their drugs reviewed for hepatotoxicity

advice to avoid alcohol until they have been seen by a hepatologist

Recommendation 21. Centres should agree monitoring, action or referral thresholds for people with CVID and abnormal LFTs with the local hepatology team.

Recommendation 22. Assess for granulomatous disease and autoimmune disease in other organs in adults with CVID and associated liver disease.

Discussion

This is the first set of UK recommendations for the monitoring and management of non‐infectious complications of CVID. We used the Delphi method to produce several consensus statements. This method was chosen as there is not a sufficiently large evidence base for the diagnosis and management of any of these complications on which to develop evidence‐based guidelines, as is often the case with rare diseases. We started with 63 statements; 59 statements reached consensus and were condensed into 22 recommendations.

The steering group identified a number of significant non‐infectious complications of CVID to address in the Delphi statements. There is an inevitable overlap with infectious disease, particularly where exclusion of infection is critical to the diagnosis of non‐infectious complications. As a result, some statements regarding infectious complications had to be included. GLILD was not addressed, as a UK consensus statement on the management of GLILD has recently been published 9.

The increased risk of malignancy in this group of patients is acknowledged; the statements recommending for regular clinical assessment (including weight) and regular monitoring blood tests were included to take this into account. Development of further guideline/consensus on screening and monitoring for cancers such as gastric cancer and lymphomas will be required in the future.

The survey respondents were representative of medical consultants specializing in immunology working in the United Kingdom, although there was a low response rate from the nurses’ group. Overall, the majority of participants (80%) had more than 10 years’ experience in immunology, cared for 50 or more CVID patients (70%) and worked in 21 of the 36 immunodeficiency centres in the United Kingdom. Fifty‐two per cent (n = 44) of the current consultant immunologist workforce contributed to developing these recommendations, including the members of the steering group.

In all the areas of the survey, respondents emphasized the importance of the multi‐disciplinary team approach for the management of patients, including expert histopathologists where appropriate. This is line with the recently published guideline on GLILD, another non‐infectious complication of CVID 9. The steering group acknowledges that individual patient management should also take into account other co‐morbidities and the presence of other manifestation/complications of CVID, and should be in accordance with the patient’s needs.

The survey was sent to immunologists, and not the organ‐based specialists who are also involved in the care of these patients. This was due to the difficulty in identifying organ‐based specialists with sufficient expertise and is a limitation of this work. However, recent UK census data show that 50% of immunology services in the United Kingdom hold combined respiratory/immunodeficiency clinics and 25% combined gastroenterology/immunodeficiency clinics (data from 2017 UK Primary Immunodeficiency Network/Royal College of Physicians Quality in Primary Immunodeficiency Services census). It is therefore probable that the responses are likely to be congruent with recommended practice in these multi‐disciplinary team (MDT) settings.

The steering group recognizes the flaws of the Delphi method, including the arbitrarily set level of consensus, the risk of bias when responding to feedback from participants and the difficulties when administrating the Delphi electronically16. Despite these, the Delphi method is useful for establishing consensus in rare disease management.

We hope that this guideline will be widely adopted and will be used by centres caring for CVID patients in the United Kingdom in order to compare their current practices with these recommendations. It will also raise further research questions and encourage audit and collaboration to address the uncertainties in several areas.

Lessons learnt

Establishing a Delphi consensus in rare diseases with few specialists can be challenging when 20% of the available medical respondents comprise the steering group.

In total, 44 consultant immunologists and four specialist nurses were involved in developing these recommendations. It is reassuring that a high level of consensus was reached in the first round, suggesting a good degree of nationwide consensus among respondents, even in the absence of high‐quality evidence or previous guidelines.

Strategies to improve input organ‐based specialists with relevant experience in pelvic inflammatory disease (PID) complications should be developed in future.

Responses from the nursing group were much lower: this may reflect a perceived lack of expertise in the diagnosis and medical management of the CVID complications described.

We cannot presuppose the reasons for non‐responses as the reasons for not completing an online survey will vary, and it is possible that agreed single individuals may have responded on behalf of larger centres). However, it will be important to explore ways to optimize response rates with the Delphi process for future surveys. This process has emphasized the need to promote awareness of these complications and provide guidance on their identification and management.

Areas for research

The steering group identified the following areas for future research.

Better description of the natural history of progression of non‐infectious complications and combinations of these patient cohorts. This will help identification and early intervention where appropriate. Next‐generation sequencing may help to categorize subgroups with predisposition for certain complications within the CVID cohort.

Improved identification of risk factors for these complications, including demographics, family history, concurrent diagnoses and genetics. This may eventually lead to pre‐emptive treatment, including targeted therapies, gene therapy and stem cell transplantation.

Increased collection of prospective data with regard to specific complications of CVID. The study of interstitial lung disease in primary antibody deficiency17 is an example of this. Similar initiatives are therefore urgently needed to improve treatment outcomes for other life‐limiting complications of CVID. This will require co‐ordinated and sustained efforts of the whole immunology community to support such prospective studies.

Conclusions

We present the first set of UK recommendations for the monitoring and management of non‐infectious complication of CVID. This will help to standardize practices and facilitate future audit, quality improvement work and research for this cohort of patients.

Disclosures

T. G. has received consultancy, lecture fees and travel support from CSL and Shire. A. H. has received lecture fees from Biotest, Shire and Octapharma and has received travel support from BPL, CSL Behring and Biotest. S. J. reports grants, personal fees, non‐financial support and other from CSL Behring, Baxalta, Weatherden and Zarodex; grants, personal fees, and non‐financial support from Shire and Biotest; grants from Binding Site; grants and personal fees from Swedish Orphan Biovitrum; personal fees and non‐financial support from Grifols, Octapharma and UCB Pharma; personal fees from LFB Group and Sanofi; and non‐financial support from Bio Products Laboratory Ltd (BPL), outside the submitted work. Y. K. has received grant research support from Roche. R. J. has received funding from Shire, Biotest, CSL and Novartis for attending conferences and symposiums and for carrying out research during his work from Octapharma, Shire and Biocryst. S. S. has received consultancy, lecture fees and travel support from CSL, Shire, Octopharma, Sobi and Novartis. Grants from CSL, Novartis, Sobi and Octopharma. D. G.‐K. has received lecture fees, advisory board and travel expenses support from CSL Behring, Shire, Grifols, Biotest and BPL. S. G. has received educational support to attend symposia by CSL Behring and Biotest. Her department received funds for research from Shire. C. B., W. E. and K. K. have no conflicts of interest to declare.

Table A1

Delphi statements where consensus was not reached in the first round on the monitoring of non‐infectious complications

	Results %
	Strongly disagree/ disagree	Strongly agree/agree
Adults with CVID should have an ultrasound of upper abdomen performed at least annually to monitor for the development of hepatosplenomegaly	64	36
Adults with CVID should have annual screening for Helicobacter pylori	53	47
Adults with CVID should not have blood tests for B12, folate and ferritin unless anaemia is present	35	65
In adults with CVID and diarrhoea for more than two weeks the following opportunistic infections should be excluded:
Human immunodeficiency virus (HIV)	40	60
Epstein–Barr virus (EBV)	39	61
Alcohol and acid‐fast bacilli (AAFB)	43	57

CVID = common variable immunodeficiency.