Abstract
Context
Lymphocytic hypophysitis (LyHy) is characterized by inflammation of the pituitary and or neuroinfundibulum and is uncommon. Treatment options include observation, high-dose glucocorticoids (HD-GCs) or surgery. Optimal first-line management strategy, however, remains unknown.
Objective
This work aimed to assess response to first-line treatment options (observation, HD-GCs, or surgery) of clinically relevant outcomes (symptomatic, hormonal, and radiographic improvement) among patients with LyHy.
Methods
A systematic review was conducted in 6 databases through 2020. Meta-analysis was conducted when feasible using a random-effects model.
Results
We included 33 studies reporting on 591 patients (423 women, 72%) with LyHy. Improvement/resolution of anterior pituitary dysfunction was highest when HD-GCs was first-line treatment. Surgery was associated with the greatest proportion of patients who had regression on imaging. Subgroup analysis comparing HD-GCs to observation showed the odds of anterior pituitary hormone recovery (OR 3.41; 95% CI, 1.68-6.94) or radiographic regression (OR 3.13; 95% CI, 1.54-6.36) were higher with HD-GCs, but so was the need for additional forms of treatment (OR 4.37; 95% CI, 1.70-11.22). No statistically significant difference was seen in recovery of diabetes insipidus (OR 0.9; 95% CI, 0.26-3.10). Certainty in these estimates was very low.
Conclusion
Observation and use of HD-GCs both are successful first-line management strategies in LyHy. Although use of HD-GCs was associated with increased recovery of anterior pituitary hormone deficit, it also was associated with greater likelihood of additional treatment after withdrawal. Optimal dosing and duration of HD-GCs remains unknown.
Keywords: adenohypophysitis, infundibuloneurohypophysitis, panhypophysitis, autoimmunity, pituitary, treatment
The umbrella term of hypophysitis may be subdivided according to etiology into primary (autoimmune) or secondary hypophysitis. Histopathologically, lymphocytic hypophysitis (LyHy), granulomatous hypophysitis, necrotizing hypophysitis, and mixed types belong to the spectrum of primary hypophysitis with suspected autoimmune background. Whether these subtypes represent the same disease or separate entities is not clear. Xanthomatous and immunoglobulin G 4 (IgG4) hypophysitis has been historically considered a manifestation of autoimmune hypophysitis. However, recent evidence suggests that xanthomatous hypophysitis rather belongs to secondary hypophysitis as it is typically caused by a Rathke cleft cyst (1). Additionally, IgG4-related hypophysitis may be a manifestation of a systemic condition (2).
Primary autoimmune hypophysitis is rare, with an estimated prevalence between 0.24% and 0.87% among lesions found during pituitary surgery and an incidence of 1 in 9 million, although this is believed to be an underestimate (3-7). A female to male predominance has been observed (8). Although antipituitary antibodies have been detected in autoimmune hypophysitis, their use as adjunct diagnostic markers has been limited by poor sensitivity and specificity of assays (9).
LyHy is the most common subtype of primary, autoimmune hypophysitis and is characterized by a dense lymphocytic inflammatory infiltrate in the adenohypophysis, neuroinfundiulum, or both with occasional plasma cells and eosinophils and in some cases lymphoid follicles (10, 11). Depending on the affected structures, LyHy is subdivided into lymphocytic adenohypophysitis, lymphocytic infundibuloneurohypophysitis, and lymphocytic panhypophysitis. Although a conclusive diagnosis is made histopathologically, a presumptive diagnosis after the exclusion of other etiologies is often feasible with appropriate clinical, biochemical, and radiological assessment (12).
As a result of inflammation, LyHy may present with clinical symptoms or signs caused by one or more pituitary hormone abnormalities and/or mass-related neurological sequelae. Headache and visual disturbances are reported to be among the most common presenting symptoms (8). Patients may present with hyperprolactinemia and various pituitary deficits (7, 8, 13-18). Consequently, the clinical spectrum of LyHy is wide and ranges from mild to severe symptoms, including death (11). As such, recognition and appropriate management are important.
The best management strategy in LyHy remains unclear. Treatment goals are to replace or prevent pituitary hormonal deficits and minimize mass effect by either reducing inflammation through immune-modulating therapies (glucocorticoids [GC] and GC-sparing agents such as azathioprine, methotrexate, and more recently rituximab) or by decompressive surgery, and rarely radiotherapy (19-22). However, given the rarity and heterogeneity of the condition and the retrospective nature of the majority of studies of small sample size, there is little information to guide optimal treatment plans. Moreover, recently published studies report conflicting clinical outcomes regarding the use of high-dose GCs (HD-GCs) in LyHy (23, 24). Some studies suggest that HD-GC therapy may lead to pituitary hormonal axis recovery (24-26) and rapid decrease in mass effect, whereas in other studies pituitary hormonal axis recovery was not consistently seen with HD-GC therapy, which can pose significant adverse effects (11, 22). Furthermore, differing doses and types of GCs have been used for variable duration. The wide variety of treatment regimens reflects the lack of consensus regarding ideal treatment strategy. Therefore, the aim of this study was to assess the optimal first-line treatment strategy (observation, HD-GCs, or surgery) associated with the greatest improvement in clinical outcomes (hormonal, radiological, or symptomatic improvement) in LyHy.
Materials and Methods
This systematic review is reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA). The protocol is registered with PROSPERO (No. CRD4201918989).
Search Strategy
A comprehensive search of several databases from each database’s inception to September 4, 2020, in any language was conducted. The databases included Ovid MEDLINE and Epub Ahead of Print, In-Process & Other Non-Indexed Citations, and Daily, Ovid EMBASE, Ovid Cochrane Central Register of Controlled Trials, Ovid Cochrane Database of Systematic Reviews, Scopus, and Web of Science. The search strategy was designed and conducted by an experienced librarian with input from the study’s principal investigator (I.B.). A controlled vocabulary supplemented with keywords was used to search for studies of autoimmune lymphocytic hypophysitis (Supplementary appendix 1) (27). Manual review of the reference list of all retrieved articles was performed to identify additional references that may have been overlooked in the search strategy.
Inclusion Criteria
We included studies in any language that included 5 or more patients with pathologically confirmed or presumed primary LyHy (based on the presence of hormonal abnormality [hypopituitary/hyperprolactinemia/diabetes insipidus; DI] and/or typical symptoms [headache or visual abnormalities] and/or radiological findings [gland/stalk enlargement, loss posterior bright spot, asymmetry] and following the exclusion of other causes) and that reported on one or more outcomes (hormonal, radiological, or symptomatic improvement) following either observation, HD-GCs, or surgical intervention. Studies solely reporting on IgG4, granulomatous, or xanthomatous hypophysitis were excluded.
Treatment groups were defined as follows: 1) the observation group included patients described as having no hormonal replacement or hormonal replacement for hypopituitarism without the use of supraphysiological or immunosuppressive doses of glucocorticoids, 2) the HD-GC group included patients who received supraphysiological doses of glucocorticoids, and 3) the surgery group included those who had surgery with the intent of resection. Symptoms, radiographic appearance, or pituitary hormonal deficits were considered improved or resolved when the authors documented improvement or resolution. In addition, pituitary hormonal deficit improvement or resolution was also considered to have occurred when hormonal replacement was no longer needed, or deficit had improved on repeat testing.
Considering the paucity of available evidence, we included the following:
1) comparative study designs (ie, studies in which patients received > 1 of the 3 interventions, thus, allowing comparison of outcomes between the interventions, regardless of randomization or other design features), and
2) noncomparative designs (ie, studies in which patients received a single intervention; thus, these studies could have a cohort or a case series design).
Study Selection
Reviewers working independently and in duplicate (Z.S. and D.D.) screened the search result titles and abstracts relevant to the research question and then performed a full-text assessment of manuscripts for eligibility. Discrepancies in study selection were resolved through discussion and consensus with a third reviewer (D.Do.). Review articles, commentaries, letters, and meeting abstracts were excluded.
Data Extraction
Working independently and in duplicate, each abstracter (Z.S. and D.D.) collected data using a standardized data extraction sheet (Microsoft Excel). Patient variables (age, sex, history of autoimmune disease, antibody testing and positivity, symptoms at presentation, testing performed), treatment received (surgery, observation, or immunosuppression), outcomes (clinical, hormonal, and radiological), and relapse were extracted when available. To homogenize the data, where available, cases reported as IgG4, granulomatous, or xanthomatous were excluded.
Methodological Quality
The methodological quality of each study was assessed in duplicate by 2 investigators (D.D. and Z.S.), and any discrepancies were solved by the third investigator (D.Do.). Methodological quality of noncomparative studies was assessed as previously described (28). For the assessment of comparative observational studies, the modified Newcastle-Ottawa tool (29) was used and assessed 1) the representativeness of patients with LyHy, 2) the assessment of outcome at baseline and follow up, and 3) the duration and adequacy of the follow-up (Supplementary appendix 2) (27).
Statistical Analysis
For comparative studies, we extracted a 2 × 2 table from each study and estimated odds ratios (OR) and 95% CIs. For noncomparative assessment that also included cohorts from comparative studies, we extracted a numerator and denominator from each study and estimated the event rate at the longest follow-up and 95% CI with variance estimated using the Freeman-Tukey double-arcsine method (30). Estimates were pooled across studies using the random-effects model because of observed heterogeneity of study populations and settings. Analysis was stratified based on intervention type (surgery, GCs, or observation).
Analysis was conducted using STATA software (StataCorp. 2019. Stata Statistical Software: release 16. StataCorp LLC.). Network meta-analysis was not feasible because the studies were not randomized and were retrospective. Comparative inferences are made only based on the retrospective comparative series that provided a 2 × 2 table. Certainty in the comparative estimates was assessed using the GRADE approach (Grading of Recommendations, Assessment, Development, and Evaluation) (31).
Results
Studies
The initial literature search revealed 434 articles; however, after initial screening 327 were excluded as outlined in the PRISMA flowchart (Fig. 1). Following the review of full-text articles, 74 studies were excluded. The most common reason for exclusion was lack of follow-up or patient data (n = 27). Ultimately, 33 studies in total were included.
Figure 1.
Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) flowchart of the study selection.
Study Characteristics
In total 33 studies representing 32 patient cohorts spanning 15 countries and including 591 unique patients with presumed (n = 335, 57%) or histologically confirmed (n = 256, 43%) LyHy were deemed eligible for inclusion (Table 1). Twenty-eight studies were comparative (6, 7, 13-18, 22-26, 32-43, 11, 44, 45) and 4 were noncomparative (46-49). A single patient cohort assessed by Honegger et al was reported in 2 different manuscripts, 1 describing presentation (50) and the other reporting treatment outcomes (22). All studies were observational case series/cohorts with only 1 study being prospective in design (24), whereas the remaining were retrospective multicenter (n = 8) or single-center (n = 24) studies. Studies were conducted in Europe (n = 14), Asia (n = 9), and North and South America (n = 8), with one study being multinational (33). Most studies involved primarily adults (mean age 40.3 ± 6.2 years; range, 25-53 years), except for one study that exclusively examined central DI due to infundibuloneurohypophysitis in children (42). Overall, 423 (72%) patients were women. The relation of hypophysitis onset to pregnancy was discussed in all but 3 studies and occurred in 84 of 416 of all women (20%), with 32 cases occurring during pregnancy (Table 2). The occurrence of LyHy in relation to timing of pregnancy was discussed in 68 cases, with 32 cases occurring during pregnancy and 37 cases occurring post partum.
Table 1.
Overview of included studies
| Author, y | Period of enrollment | Country | Study type | Total No., n | Female n, % | Age at presentation, y | Diagnosis, n, % | Therapy | Follow-up, mo |
|---|---|---|---|---|---|---|---|---|---|
| Imura, 1993 (37) | 1986-1992 | Japan | Retrospective, single center | 17 | 13 (76%) | Mean: 48 ± 9.1 | Clinical: 15, 88%; histological: 2, 12% | Observation: 15, 88%; HD-GCs: 2, 12% | NR |
| Ahmadi, 1995 (32) | 1989-1994 | USA | Retrospective, multicenter | 5 | 4 (80%) | Mean: 25 ± 15.8 | Clinical: 1, 20%; histological: 4, 80% | Surgery: 2, 40%; HD-GCs: 3, 60% | Range: 18-36 |
| Thodou, 1995 (11) | NR | Canada | Retrospective, single center | 16 | 14 (88%) | Mean: 32.7 ± 6.8 | Clinical: 0; Histological: 16, 100% | Observation: 3, 19%; surgery: 12, 75%; HD-GCs: 1, 6% | NR |
| Honegger, 1997 (6) | 1982-1995 | Germany | Retrospective, single center | 6 (total 9, excluded 3 GrHy) | 4 (67%) | a Mean: 33.5 ± 9 | Clinical: 0; histological: 6, 100% | Surgery: 4, 67%; HD-GCs: 2, 33% | Mean: 40.2 ± 33.3 |
| Crock, 1998 (33) | NR | Canada, USA, and Australia | Retrospective, multicenter | 32 | 23 (72%) | Mean: 39.5 ± 14 | Clinical: 22, 69%; histological: 10, 31% | Observation: 24, 75%; surgery: 8, 25% | Range: 3-84 |
| Kristof, 1999 (48) | NR | Germany | Retrospective, single center | 9 | 7 (78%) | Mean: 40.9 ± 13.7 | Clinical: 7, 78%; histological: 2, 22% | Surgery: 1, 11%; HD-GCs: 8, 88% | Mean: 28.9 ± 9, range 19-42 |
| Flanagan, 2002 (34) | 1985-2002 | UK | Retrospective, single center | 8 (total 14, excluded 2 GrHy, 4 RCC) | 7 (88%) | Mean: 53.1 ± 14.5 | Clinical: 2, 25%; histological: 6, 75% | Observation: 2, 25%; surgery: 6, 75% | Mean: 53 ± 14.5 |
| Leung, 2004 (7) | 1992-2003 | USA | Retrospective, single center | 13 (total: 16, excluded GrHy) | 7 (54%) | a Mean: 46.2 ± 15.6 | Clinical: 3, 23%; histological: 10, 77% | Observation: 8, 62%; HD-GCs: 5, 38% | Mean: 24.8, range: 4-48 |
| Gutenberg, 2006 (16) | 1992-2005 | Germany | Retrospective, multicenter | 21 (total 31, excluded 6 GrHy, 4 XaHy) | 10 (48%) | Mean: 43.2 | Clinical: 0; histological: 21, 100% | Surgery: 21, 100% | Median: 24, range: 1-133 |
| Menon, 2009 (13) | 1988-2006 | India | Retrospective, single center | 15 | 14 (93%) | a Mean: 37 ± 10 | Clinical: 12, 80%; histological: 3, 20% | Observation: 12, 80%; surgery: 3, 20% | Mean: 15.8, range: 3-43 |
| Nakata, 2010 (49) | 1987-2009 | Japan | Retrospective, Multicenter | 20 | 10 (50%) | b Mean: 43 ± 21 | Clinical: 16, 80%; histological: 4, 20% | Observation: 9, 45%; surgery: 11, 55% | Mean: 32.3 ± 30.5 |
| Iuliano, 2011 (47) | NR | USA | Retrospective, single center | 5 (total 7, 2 excluded germinoma and GrHy) | 3 (60%) | a Mean: 48.2 ± 14.7 | Clinical: 0; histological: 5, 100% | Surgery: 5, 100% | NR |
| Pérez, 2013 (41) | 2000-2012 | Mexico | Retrospective, single center | 9 | 7 (78%) | a Mean: 34 ± 12 | Clinical: 9, 100%; histological: 0 | Observation: 4, 44%; HD-GCs: 5, 56% | NR |
| Park, 2014 (40) | 2001-2003 | Korea | Retrospective, single center | 22 (including 2 GrHy and 1 XaHy) | 17 (77%) | Mean: 48; range 21-67 | Clinical: 11, 50%; histological: 11, 50% | Observation: 12, 54%; surgery: 5, 23%; HD-GCs: 5, 23% | Mean: 57; range 7-138; median: 48 |
| Heaney, 2015 (46) | 1995-2012 | USA | Retrospective, single center | 15 | 11 (73%) | Mean: 44.8 ± 12.6 | Clinical: 3, 20%; histological: 12, 80% | Unclear which type treatment given first | NR |
| Honegger, 2015 (50) | 2000-2013 | Germany | Retrospective, multicenter | 76 (including 9 GrHy) | 54 (71%) | Mean: 41.0; range 16-81 | Clinical: 40, 53%; histological: 36, 47% | Observation: 30, 39%; surgery: 19, 25%; HD-GCs: 23, 30%; other: steroid + surgery, 3, 4% immunosuppressant: 1, 1% | Mean: 45.6, median: 22.8 |
| Imber, 2015 (36) | 1997-2014 | USA | Retrospective, single center | 16 (total 21, exclude 4 GrHy, 1 IgG4, 1 mixed) | 10 (63%) | a Mean: 39 ± 11 | Clinical: 2, 13%; histological: 14, 87% | Surgery: 3, 19%; HD-GCs: 13, 81% | a Mean: 39.8 |
| Khare, 2015 (14) | 2006-2012 | India | Retrospective, single center | 24 | 21 (88%) | Mean: 31.5 ± 9.57 | Clinical: 19, 79%; histological: 5, 21% | Observation: 15, 63%; surgery: 5, 21%; HD-GCs: 4, 16% | Median: 12 |
| Tauziede-Espariat, 2015 (43) | 1993-2013 | France | Retrospective, single center | 5 (Total 8, excluded 2 IgG4, I GrHy) | 5 (100%) | a Mean: 41 ± 16.4 | Histological: 5, 100% | Observation: 3, 60%; HD-GCs: 2, 40% | NR |
| Guo, 2016 (35) | 2009-2011 | China | Retrospective, single center | 5 (total 7, excluded 2 GrHy) | 3 (60%) | a Mean: 39.4 ± 23.5 | Clinical: 2, 40%; histological: 3, 60% | Observation: 1, 20%; surgery: 1, 20%; HD-GCs: 3, 60% | Median: 24 |
| Tirosh, 2016 (44) | 2008-2014 | Israel | Retrospective, single center | 9 | 9 (100%) | Mean: 33.7 ± 7.8 | Clinical: 9, 100%; | Observation: 7, 78%; HD-GCs: 2, 22% | Median: 23, range 7-67 |
| Kyriacou, 2017 (15) | 1990-?2016 | UK | Retrospective, Multicenter | 22 | 19 (86%) | a Mean: 39.7 ± 13.2 | Clinical: 14, 64%; histological: 8, 36% | Observation: 12, 55%; surgery: 8, 36%; HD-GCs: 2, 9% | Mean: 103.2, range: 60-312 |
| Lupi, 2017 (23) | 2008-2016 | Italy | Retrospective, single center | 11 (including 1 GrHy) | 8 (73%) | a Mean: 47.7 ± 14.2 | Clinical: 10, 99%; histological: 1, 1% | Observation: 4, 36%; HD-GCs: 7, 64% | Mean: 30 ± 27 |
| Schaefers, 2017 (42) | 2002-2017 | Belgium | Retrospective, Multicenter | 9 | 5 (56%) | Median: 6.7, range: 3-14 | Clinical: 9, 100%; histological: 0 | Observation: 8, 89%; steroid therapy: 1, 11% | Mean: 4.8, range: 18-60 |
| Wang, 2017 (18) | 1999-2016 | China | Retrospective, single center | 50 | 33 (66%) | Mean: 37.2 ± 11.2 | Clinical: 28, 56%; histological: 22, 44% | Observation: 9, 18%; surgery: 15, 30%; HD-GCs: 26, 52% | NR |
| Angelousi, 2018 (26) | 1993-2015 | Greece | Retrospective, multicenter | 22 (including 1 XaHy, 1 GrHy) | 17 (77%) | Mean: 41.9 ± 11.8 | Clinical: 8, 36%; histological: 14, 64% | Observation: 5, 23% surgery: 9, 41%; HD-GCs: 8, 36% | Median: 48, IQR 3-60 |
| Chiloiro, 2018 (24) | 2008-2015 | Italy | Prospective, single center | 20 | 14 (70%) | Median (IQR): steroid arm: 42 (29); observation arm: 36 (31.2) | Clinical: 20, 100%; histological: 0, 0% | Observation: 8, 40%; HD-GCs: 12, 60% | Median: 123, IQR 99 |
| Panigrahi, 2018 (39) | 2013-2015 | India | Retrospective, single center | 9 (total 12, excluded 3 GrHy) | 5 (56%) | a Mean: 38.44 ± 19.83 | Clinical: 5, 66%; histological: 4, 44% | Surgery: 3, 33%; HD-GCs: 6, 66% | Mean: 9.33 ± 4.4 |
| Korkmaz, 2019 (38) | 2007-2018 | Turkey | Retrospective, single center | 17 (including 1 XaHy, 1 mixed) | 10 (59%) | Median: 31 | Clinical: 9, 53%; histological: 8, 47% | Observation: 10, 59%; surgery: 7, 41% | Median: 24 (6-84) |
| Zhu, 2019 (45) | 2011-2018 | China | Retrospective, single center | 20 | 16 (20%) | Median: 37 (16-58) | Clinical: 10, 50%; histological: 10, 50% | Observation: 4, 20%; surgery: 10, 50%; HD-GCs: 6, 30% | Median: 37 (16-58) |
| Amereller, 2020 (25) | 1997-2016 | Germany | Retrospective, single center | 51 (total 60, excluding 9 in Honneger et al) | 36 (71%) | a Mean: 46 ± 16 | Clinical: 45, 80%; histological: 6, 20% | Observation: 35, 69%; surgery: 6, 12%; HD-GCs: 10, 20% | a Mean: 69 ± 75 |
| Oguz, 2020 (17) | 1999-2017 | Turkey | Retrospective, single center | 12 (total 20, excluded 6 GrHy, 1 XaHy, 1 mixed) | 7 (58%) | a Mean: 37 ± 12 | Clinical: 4, 25%; histological: 8, 75% | Observation: 1, 8%; surgery: 8, 67%; HD-GCs: 3, 25% | NA |
| Total | 1982-2018 | Europe: 14, Asia: 10, Americas: 7 , multicontinental: 1 | Retrospective: 31 studies; prospective: 1 study | 591 | 423 (72%) | Mean: 40.3 ± 6.2, range: 3-81 | Clinical: 335, 57%; histological: 256, 43% | Observation: 241 surgery: 172 HD-GCs: 159 combination/other: 4 | Mean: 40 ± 19.1, range: 1-312 |
Abbreviations: GrHy, granulomatous hypophysitis; HD-GCs, high-dose glucocorticoids; IQR, interquartile range; NA, not available; NR, not reported; RCC, renal cell carcinoma; UK, unknown; USA, United States of America; XaHy, xanthomatous hypophysitis.
a Calculated from individual patient data provided by authors or available in the manuscript.
b Calculated differently from the paper.
Table 2.
Symptoms, relation to pregnancy, and associated autoimmunity on presentation in patients with lymphocytic hypophysitis
| Author, y | Symptoms at diagnosis N/total, % | Associated with pregnancy N/No. of women, % | AI Abs, N/No. tested, % | AI conditions, N/total, type | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Headache | Visual field defects | CN palsies | Hypogonadism | Galactorrhea | Fatigue | Polyuria/polydipsia | ||||
| Imura, 1993 (37) | 0/17 | 1/17, 11% | 0/17 | 0/17 | 0/17 | 0/17 | c 17/17, 100% | 0/13b | ANA:3, RF:2 | 3/17, 18% (DM not specified: 1, glomerulonephritis:1, Graves disease:1) |
| Ahmadi, 1995 (32) | 5/5, 100% | 3/5, 60% | 1/5, 20% | 3/5, 60% | 1/5, 20% | 0/5, 0% | 0/5, 0% | 2/4, 50% | NR | NR |
| Thodou, 1995 (11) | 5/16, 31% | 7/16, 44% | 0/16 | 1/16, 6% | 0/16 | 0/16 | 3/16, 19% | 10/14, 71%, AP: 4, PP: 6 | APA: 0/2 | 4/ 16, 25% (AI thyroiditis: 4 (2 hyper, 2 hypo), vitiligo + adrenalitis:1, AI pancreatitis: 1 (also thyroiditis, parathyroid hyperplasia) |
| Honegger, 1997 (6) | 0/6 | 0/6 | 0/6 | 0/6 | 0/6 | 0/6 | 1/6, 17% | 0/4 | NR | NR |
| Crock, 1998 (33) | 9/32, 28% | 7/32, 22% | 0/32 | 4/32, 16% | 13/32, 41% | 7/32, 22% | 2/32, 6% | 10/23, 43% | APA 49kD: 19/32, APA 40 kDA: 11/32 | 13/32, 41% (IDDM: 8/32; HT: 4/32; ANF: 8/32; pernicious anemia: 1; Graves disease: 1; Addison: 1; iritis: 1) |
| Kristof, 1999 (48) | 1/9, 11% | 1/9, 11% | 0/9 | 3/9, 33% | NR | 1/9, 11% | 4/9, 44% | 1/7, 14%, AP | TgAb 1/4 | NR |
| Flanagan, 2002 (34) | 6/8, 75% | 2/8, 25% | 1/8, 13% | 1/8, 13% | 2/8, 25% | 2/8, 25% | 2/8, 25% | 1/7, 14% AP | NR | 2/8, 25% (hypothyroidism and RA: 1, Crohn disease: 1) |
| Leung, 2004 (7) | 9/13, 69% | 4/13, 31% | 0/13 | 9/13, 69% | 1/13, 8% | 9/13, 69% | 3/13, 23% | 3/7, 43% (1 AP, 2 PP) | NR | 5/13, 38%, (hypothyroidism unknown nature: 1, temporal arteritis: 1, Graves disease: 1, orbital pseudotumor: 1, idiopathic myopathy: 1) |
| Guttenberg, 2006 (16) | 14/21, 67% | 9/21, 43% | 0/21 | 11/21, 52% | NR | 0/21 | 10/21, 48% | 3/10, 30%, third trimester | NR | 5/21, 24% (GPA:1, psoriasis: 1, T1DM: 1, HT: 1, polymyositis: 1) |
| Menon, 2009 (13) | 10/15, 67% | 4/15, 27% | 2/15, 13% | 4/15, 27% | 0/15 | 6/15, 40% | 1/15, 7% | 3/14, 21%, 2AP, 1 PP | NR | 2/15, 13% (Sjögren syndrome: 1, RA: 1, AI thyroid disease: 1 |
| Nakata, 2010 (49) | 6/20, 30% | 2/20, 10% | 1/20, 5% | 4/20, 20% | NR | 3/20, 15% | 16/20, 80% | 1/10, 10%, AP | NR | 5/20, 25% (RA: 1, bullous pemphigoid: 1, AI pancreatitis: 2, AI hepatitis: 1, interstitial nephritis: 1, chronic thyroiditis: 1) |
| Iuliano, 2011 (47) | 4/5, 80% | 0/5 | 0/5 | 3/5, 60% | 1/5, 20% | 3/5, 60% | 2/5, 40% | 1/3, 33% (1-y PP) | NR | 3/5, 60%; (HT: 2, Grave disease: 1) |
| Perez, 2013 (41) | 2/9, 22% | NR | NR | 3/9, 33% | NR | NR | 7/9, 78% | 0/7 | TPO Ab: 3/9, 33% | 3/9, 33% (HT: 1, Graves disease: 1, antiphospholipid syndrome:1) |
| Park, 2014 (40) | 6/22, 27% | 2/22, 9% | NR | 3/22, 14% | NR | NR | 18/22, 82% | 1/17, 6%, AP | NR | 2/22, 9%, (HT: 1, GPA: 1) |
| Heaney, 2015 (46) | 10/15, 67% | 5/15, 30% | 0/15 | 11/15, 73% | 0/15 | 0/15 | 8/15, 53% | NR | NR | 2/15, 13%; (HT:1, polyendocrine AI syndrome: 1) |
| Honegger, 2015 (50) | 38/71, 54% | 11/76, 14% | 5/76, 7% | NR | NR | NR | NR | 6/54, 11% | TPO Ab: 28/76, 37% | 12/76, 16%, (HT: 7, Graves disease: 3, Crohn disease: 1, Sjögren syndrome: 1) |
| Imber, 2015 (36) | 9/16, 56% | 8/16, 50% | NR | 8/16, 50% | 1/16, 6% | 6/16, 38% | 9/16, 56% | 4/10, 40% (1 AP, 2 PP, 1 not specified) | NR | 2/21, 10% (T1DM: 1, thyroid: 1) |
| Khare, 2015 (14) | 20/24, 83% | 2/24, 8% | 3/20, 15% | 10/24, 42% | 3/24, 13% | NR | 4/24, 17% | 1/20, 5%, AP | NR | NR |
| Tauziede-Espariat, 2015 (43) | NR | NR | NR | NR | NR | NR | 2/5 | 1/5, 20%, PP | NR | NR |
| Guo, 2016 (35) | 3/5, 60% | 1/5, 20% | 0/5 | 0/5 | 0/5 | 0/5 | 0/5 | 1/5, 20%; AP | NR | NR |
| Tirosh, 2016 (44) | 5/9, 56% | 0/9 | NR | 8/9, 89% | NR | 4/9, 44% | 1/9, 11% | 9/9, 1 AP, 8 PP | NR | NR |
| Kyriacou, 2017 (15) | 15/22, 68% | 7/22, 32% | 0/22 | 6/22, 27% | 1/22, 5%, (4/22 failure to lactate) | 4/22, 18% | 1/22, 5% | 11/19, 58%, (5 AP, 6 PP) | NR | 4/24, 17% (AI hypothyroidism: 2, RA: 1, premature ovarian insufficiency: 1) |
| Lupi, 2017 (23) | 3/11, 27% | NR | NR | 4/11, 36% | NR | 4/11, 36% | 9/11, 82% | NR | 6/11, 55%, APA | NR |
| Schaefers, 2017 (42) | 1/9, 11% | 0/9 | 0/9 | NR | 0/9 | 0/9 | 9/9, 100% | NR | 1/2, 50%, APA, ANCA, anti-SMA, ANA | None |
| Wang, 2017 (18) | 24/50,48% | 20/50, 40% | 0/50 | NR | NR | NR | NR | 8/33, 24% | NR | 3/50, 6%; HT: 2, Graves disease: 1 |
| Angelousi, 2018 (26) | 1/22, 59% | 7/22, 32% | NR | 13/22, 59% | NR | NR | NR | 0/17 | APA: 1/10, 10%, | 8/22, 36% (thyroiditis: 3, RA: 3, vitiligo: 1, GPA: 1) |
| Chiloiro, 2018 (24) | NR | NR | NR | NR | NR | NR | NR | 0/14 | APA: 14/20, 70%; AHA: 17/20, 85% | NR |
| Panigrahi, 2018 (39) | 8/9, 89% | 2/9, 22% | 2/9, 22% | 2/9, 22% | NR | 2/9, 22% | 0/9 | 0/5 | ANA: 1 | None |
| Korkmaz, 2019 (38) | 9/17, 53% | 2/17, 12% | NR | 3/17, 18% | 1/17, 6% | 7/17, 41% | 8/17, 47% | 0/10 | NR | NR |
| Zhu, 2019 (45) | 8/20, 40% | 10/20, 50% | NR | 5/20, 25% | 2/20, 10% | 1/20, 5% | 11/20, 55% | 4/16, 25% (1 AP, 3 PP) | 0/10 (ssA, ssB, TPO, TRAb, and TgAb) | None |
| Amereller, 2020 (25) | 20/51, 39% | 7/51, 14% | 1/51, 2% | 14/51, 27% | 0/51 | 27/51, 53% | 19/51, 37% | b 2/44, 5% (1 AP, 1 PP) | NR | b 25/60, 42% (20 AI thyroiditis) |
| Oguz, 2020 (17) | 7/12, 58% | 3/12, 25% | 0/12 | 6/12, 50% | NR | 1/12, 8% | 5/12, 42% | b 1/15, 7% | 1/6 | NR |
| Totals | 253/561, 43% | 127/546, 23% | 16/436, 4% | 139/431, 32% | 26/296, 9% | 87/363, 24% | 172/423, 41% | 84/416, 20% | ||
Abbreviations: Ab, antibody; ACTH, adrenocorticotropin; AHA, antihypothalamic antibodies; AI, autoimmune; ANA, antinuclear antibodies; ANCA, antineutrophil cytoplasmic antibodies; ANF, antinuclear factor; anti-SMA, anti-smooth muscle; AP, antepartum; APA, antipituitary antibodies; CN, cranial nerve; DI, diabetes insipidus; DM, diabetes mellitus; GH, growth hormone; GPA, granulomatous with polyangiitis; HT, Hashimoto thyroiditis; IDDM, insulin-dependent diabetes mellitus; PP, postpartum; PRL, prolactin; NR, not reported or not done; RA, rheumatoid arthritis; T1DM, type 1 diabetes mellitus; TgAb, thyroglobulin autoantibodies; TPO, thyroid peroxidase; TRAb, thyrotropin receptor antibodies; TSH, thyrotropin.
a Excluding those in whom it was deemed physiological (pregnant).
b Based on the original paper’s total.
c DI was a specification for inclusion.
Baseline Presentation
The most common presenting symptom was headache (253/561, 43%) followed by polyuria/polydipsia (172/423, 41%), and symptoms consistent with hypogonadism (139/431, 32%) (see Table 2). Antibody assessment, in particular antipituitary antibodies, was reported by only some studies and when reported assessment was not performed in all patients (23, 24, 26, 33, 42, 11). An additional autoimmune condition was reported in the studies in 6% to 60% of patients, with autoimmune thyroid disease (Hashimoto thyroiditis or Graves disease) being the most prevalent.
Following hormonal evaluation, gonadotrophin deficiency (304/53, 54%) was the most common anterior pituitary hormonal deficit identified on presentation, followed closely by adrenal insufficiency (297/579, 51%) (Table 3). DI was present in 274 of 583 (47%) overall but varied between 0% and 100% in the included studies. Imaging characteristics were reported in all studies, with 64% of patients demonstrating uniform contrast enhancement, 62% symmetric pituitary enlargement, and 61% pituitary stalk thickening with a minority (10%) exhibiting cavernous sinus involvement. Reported first-line treatment for LyHy included observation (n = 241, 42%), HD-GCs (n = 159, 28%), and surgery (n = 172, 30%) with 4 (0.6%) receiving combination therapy from the outset (see Table 1). HD-GC treatment regimens varied widely and included GCs such as prednisone (maximum reported 60 mg/d), methylprednisolone (maximum reported 1000 mg/d), or dexamethasone (maximum reported 1 mg/kg/d) with varying tapering strategies and therefore duration (supplementary appendix 3, Table 1) (27).
Table 3.
Pituitary deficiencies and radiological findings in patient with lymphocytic hypophysitis on presentation
| Author, y | Pituitary deficiencies at initial presentation, N/total assessed | Imaging characteristics at initial presentation, N/total assessed | Anatomic classification | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ACTH | TSH | GH | Increased PRL | GnT | DI | Uniform contrast enhancement | Symmetric enlargement | Supra-sellar extension | Thickened stalk | Loss of bright spot | Dural tail | Cavernous sinus extension | ||
| Imura, 1993 (37) | 0/15 | 0/17 | 9/15, 60% | 2/16, 13% | 0/17 | 17/17, 100% | NR | NR | NR | 7/17, 41% | 17/17, 100% | NR | NR | NR |
| Ahmadi, 1995 (32) | 2/5, 40% | 2/5, 40% | NR | 3/5, 60% | 3/5, 60% | 0/5, 0% | 2/5, 40% | NR | NR | 4/5, 80% | 3/5, 60% | 4/5, 80% | 1/5, 20% | NR |
| Thodou, 1995 (11) | 9/16, 56% | 8/16, 50% | 4/11, 36% | 4/10a, 40% | 5/12a, 42% | 3/16, 19% | NR | NR | 10/12, 83% | 2/12, 17% | NR | NR | 2/12, 17% | NR |
| Honegger, 1997 (6) | 2/6, 33% | 1/6, 17% | 3/5, 60% | 3/6, 50% | 4/6, 67% | 3/6, 50% | 4/6, 67% | NR | 3/6, 50% | 5/6, 83% | NR | NR | NR | |
| Kristof, 1999 (48) | 6/9, 67% | 2/9, 22% | 7/9, 78% | 4/9, 44% | 5/9, 56% | 4/9, 44% | 8/9, 89% | 6/9, 67% | 7/9, 78% | 6/9, 67% | 2/9, 22% | NR | NR | NR |
| Crock, 1998 (33) | 23/31, 74% | 14/31, 45% | 13/31, 42% | 4/31b, 13% | 19/20a, 95% | 5/31, 16% | NR | NR | 8/32, 25% | 2/32, 6% | NR | NR | NR | NR |
| Flanagan, 2002 (34) | 3/7, 43% | 3/8, 38% | NR | 4/5, 80% | 6/7, 86% | 3/8, 38% | NR | NR | NR | 8/8, 100% | NR | NR | NR | NR |
| Leung, 2004 (7) | 10/13, 77% | 8/12, 67% | 5/12, 42% | 4/12, 33% | 12/12, 100% | 3/13, 23% | NR | NR | 10/13, 77% | 4/13, 31% | NR | NR | 2/13, 15% | NR |
| Guttenberg, 2006 (16) | 15/21, 71% | 17/21, 81% | 3/7, 43% | 3/19, 16% | 16/20, 80% | 10/21, 48% | NR | NR | 9/16, 56% | 11/16, 69% | NR | NR | NR | NR |
| Menon, 2009 (13) | 10/15, 67% | 8/15, 53% | 0/15 | 2/15, 13% | 6/15 40% | 0/15 | NR | NR | 10/15, 67% | 5/15, 33% | NR | NR | NR | NR |
| Nakata, 2010 (49) | 6/20, 30% | 5/20, 25% | 9/20, 45% | 6/20, 30% | 12/20, 60% | 15/20, 75% | 13/19, 68% | 17/20, 85% | NR | 17/20, 85% | 18/20, 90% | 13/19, 64% | 5/20, 25% | |
| Iuliano, 2011 (48) | 1/2, 50% | 1/2, 50% | NR | 2/4, 50% | 2/2, 100% | 2/4, 50% | NR | 3/5, 60% | 2/5, 40% | 5/5, 100% | NR | NR | 2/5, 40% | NR |
| Perez, 2013 (41) | 3/9, 33% | 1/9, 11% | NR | 1/9, 11% | 3/9, 33% | 7/9, 78% | 8/9, 89% | 8/9, 89% | NR | 9/9, 100% | 7/9, 78% | NR | NR | AH (2), INH (5), PH (2) |
| Park, 2014 (40) | 10/22, 45% | 8/22, 36% | 5/22, 23% | 5/22, 23% | 7/22, 32% | 18/22, 82% | 7/22, 32% | 7/22, 32% | 7/22, 32% | 17/22, 77% | 18/22 , 82% | NR | NR | AH (3), INH (9), PH (10) |
| Imber, 2015 (36) | 10/16, 63% | 10/16, 63% | 12/16, 75% | 8/16, 50% | 13/16, 81% | 9/16, 56% | 10/15, 67% | NR | 5/15, 33% | 9/15, 60% | 6/15m, 40% | 1/15, 67% | 1/15, 67% | NR |
| Heaney, 2015 (46) | 8/15, 53% | 10/15, 67% | NR | 6/13, 46% | 11/14, 79% | 8/15, 53% | NR | 13/15, 87% | NR | 9/15, (60) | 4/15, (27) | NR | NR | NR |
| Honegger, 2015 (6) | 34/73, 47% | 35/73, 48% | 26/70, 37% | 21/58, 36% | 45/73, 62% | 39/72, 46% | 53/66, 80% | NR | 54/72, 75% | 62/72, 86% | NR | NR | 5/72, 7% | NR |
| Khare, 2015 (14) | 18/24, 75% | 14/24, 58% | NR | 10/24, 42% | 12/24, 50% | 4/24, 17% | 22/24, 92% | 22/24, 92% | 21/24, 88% | 21/24, 88% | 17/24, 71% | 2/24, 8% | NR | NR |
| Tauziede-Espariat, 2015 (43) | 5/5, 100% | 5/5, 100% | 5/5, 100% | NR | 5/5, 100% | 2/5, 40% | NR | 4/5, 80% | 2/5, 40% | NR | NR | NR | NR | NR |
| Guo, 2016 (35) | 3/5, 60% | 3/5, 60% | NR | 0/5 | 1/5, 20% | 0/5 | 5/5, 100% | 5/5, 100% | NR | 5/5, 100% | NR | NR | NR | NR |
| Tirosh, 2016 (44) | 8/9, 89% | 6/7, 86% | 8/8, 100% | 1/9, 11% | 8/9, 89% | 1/9, 11% | NR | NR | NR | NR | 1/9, 11% | NR | NR | NR |
| Kyriacou, 2017 (15) | 19/22, 86% | 13/22, 59% | 3/22, 14% | 13/20a, 65% | 9/22, 41% | 7/22, 32% | NR | NR | 12/22, 55% | NR | NR | NR | 0/22 | NR |
| Lupi, 2017 (23) | 3/11, 27% | 4/11, 36% | 5/11, 45% | 3/11, 27% | 5/11, 45% | 9/11, 82% | 3/11, 27% | 3/11, 27% | NR | 11/11, 100% | 11/11, 100% | 2/12, 17% | NR | INH (8), PH (4) |
| Schaefers, 2017 (42) | 0/9 | 1/9, 11% | 1/3, 33% | 0/9 | 0/9 | 9/9, 100% | 9/9, 100% | 9/9, 100% | 0/9 | 9/9, 100% | 9/9, 100% | 0/9 | 0/9 | INH (9) |
| Wang, 2017 (18) | 13/50, 30% | 19/50, 38% | 9/50, 18% | 24/50, 48% | 30/50, 60% | 36/50, 72% | 39/50, 78% | 44/50, 88% | 39/50, 78% | 48/50, 96% | NR | 5/50, 10% | NR | NR |
| Angelousi, 2018 (26) | 16/22, 73% | 14/22, 64% | 13/22, 59% | 8/22, 36% | 18/22, 82% | 7/22, 32% | 5/22, 23% | 5/22, 23% | 5/22, 23% | 8/22, 36% | 4/22, 18% | NR | NR | NR |
| Chiloiro, 2018 (24) | 6/20, 30% | 2/20, 10% | 4/20, 20% | 5/20, 25% | 7/20, 35% | 11/20, 55% | NR | NR | NR | 18/20, 90% | 12/20, 60% | NR | NR | NR |
| Panigrahi, 2018 (39) | 6/9, 67% | 5/9, 56% | 0/9 | 1/9, 11% | 2/9, 22% | 0/9 | 1/3, 33% | 1/3, 33% | 2/3, 67% | 2/3, 67% | 1/3, 33% | 1/3, 33% | NR | NR |
| Korkmaz, 2019 (38) | 10/17, 59% | 9/17, 53% | 1/17, 6% | 7/17, 41% | 8/17, 47% | 8/17, 47% | 3/17, 18% | NR | 2/17, 12% | 7/17, 41% | 8/17, 47% | NR | NR | NR |
| Zhu, 2019 (45) | 4/20, 20% | 5/20, 25% | 1/20, 5% | 5/20, 25% | 2/20, 10% | 11/20, 55% | 9/20, 45% | 11/20, 55% | 10/20, 50% | 9/20, 45% | 9/20. 45% | NR | NR | NR |
| Amereller, 2020 (25) | 30/51, 59% | 27/51, 53% | 5/51, 10% | 10/51, 20% | 22/51, 43% | 19/51, 37% | NR | 18/51, 35% | NR | 27/51, 53% | 2/51, 4% | NR | NR | NR |
| Oguz, 2020 (17) | 4/10, 40% | 6/10, 60% | 3/11, 27% | 2/11, 18% | 6/10, 60% | 4/10, 40% | NR | 3/11, 27% | 6/11, 55% | 3/11, 27% | 4/11, 36% | NR | NR | NR |
| Totals | 297/579, 51% | 266/579, 46% | 154/482, 32% | 171/503, 34% | 304/563, 54% | 274/583, 47% | 201/312, 64% | 179/291, 62% | 224/400, 56% | 324/534, 61% | 153/309, 50% | 28/137, 20% | 18/173, 10% | AH (5), INH (31), PH (16) |
Abbreviations: ACTH, adrenocorticotropin; AH, adenohypophysitis; AHA, antihypothalamic antibodies; AP, antepartum; APA, antipituitary antibodies; DI, diabetes insipidus; GH, growth hormone; GnT, gonadotrophin; INH, infundibuloneurohypophysitis; NR, not reported; PH, panhypophysitis; PP, post partum; PRL, prolactin; TSH, thyrotropin.
a Excluding those in whom it was deemed physiological (pregnant).
b One patient was at 28/40 weeks’ gestation.
Outcomes of Patients With Hypophysitis
Patients were followed for a mean of 40 months, ranging from 1 to 312 months. Overall, at final follow-up considering all therapies and in whom follow-up was reported, 67% had an improvement in radiological findings, 27% had anterior pituitary dysfunction, whereas only 12.7% of patients had resolution of DI.
Although patients presented with multiple and diverse symptoms attributable to mass effect and/or hormonal deficiencies, follow-up of these symptoms was infrequently reported except for headache or visual symptoms and were seldom reported according to treatment modality (Supplementary appendix, Table 2) (27). Therefore, only outcomes regarding headaches or visual symptoms were assessed in this study. Irrespective of treatment employed, 74% (110/148) of patients were noted to have an improvement in headache and 77% (99/129) had an improvement in visual symptoms including vision, visual field abnormalities, or ophthalmoplegia. Among those who had surgery, time to headache improvement was almost immediate.
Noncomparative Series
The proportion of patients who had an improvement in anterior pituitary deficits (partial or full recovery) or recovery of DI following observation, HD-GC, or surgery is presented in Figs. 2 and 3. Interestingly, patients were less likely to develop 1 or more new hormonal deficits with HD-GC treatment (observation 19/203, 9.3%; HD-GC 6/140, 4.2%; or surgery 16/115, 13.9%). The proportion of patients who demonstrated progression or relapse and required additional treatment is presented in Fig. 4 and was highest in those who received HD-GCs. In the event of relapse or progression, second-line treatment included surgery, HD-GCs, or less commonly alternative immunosuppressive regimens such as methotrexate, azathioprine, or rituximab and occasionally radiotherapy. The time to relapse was not consistently reported but ranged from 3 months to 16 years. With regard to radiological findings, the greatest proportion of patients had an improvement in radiological findings following surgery (Fig. 5).
Figure 2.
Proportion of patients who had recovery or improvement of anterior pituitary hormone deficits according to treatment. Forest plot showing the proportion of patients who had recovery or improvement at last follow-up of anterior pituitary hormone deficits noted at diagnosis according to the initial treatment received. HD-GCs, high-dose glucocorticoids.
Figure 3.
Proportion of patients who had recovery or improvement of diabetes insipidus. Forest plot showing the proportion of patients who had recovery or improvement at last follow-up of diabetes insipidus noted at diagnosis according to the initial treatment received. HD-GCs, high-dose glucocorticoids.
Figure 4.
Proportion of patients who demonstrated progression or relapse and required additional treatment. Forest plot showing the proportion of patients who demonstrated progression or relapse and required additional treatment. HD-GCs, high-dose glucocorticoids.
Figure 5.
Proportion of patients who had recovery or improvement of image findings. Forest plot showing the proportion of patients who had recovery or improvement at last follow-up of image findings noted at diagnosis according to the initial treatment received.
Comparative Studies
Among those who at baseline presented with pituitary hormone deficits, the odds of anterior pituitary axis recovery or improvement increased with the use of HD-GCs (OR 3.41; 95% CI, 168-6.94) when compared to observation (Fig. 6A; Table 4). In contrast, no difference in recovery from DI was seen when HD-GCs were used compared to observation (OR 0.90; 95% CI, 0.26-3.10) (Fig. 6B). Moreover, compared to observation, those who received HD-GCs as first-line treatment were more likely to require additional treatment (OR 4.37; 95% CI, 1.7-11.22).
Figure 6.
Outcome comparison of high-dose glucocorticoids (HD-GCs) vs observations. Forest plots comparing A, improvement or recovery of anterior pituitary hormone deficits; B, improvement or regression of diabetes insipidus; and C, improvement or regression of image findings in those treated with HD-GCs as first-line therapy compared to those observed.
Table 4.
Grading of Recommendations, Assessment, Development, and Evaluation summary of findings
| Population: patients with lymphocytic hypophysitis Intervention: high-dose glucocorticoid Comparison: observation |
|||
|---|---|---|---|
| Outcome | No. of studies and participants | Treatment effect | Certainty in the evidence |
| Anterior pituitary hormone recovery/improvement | 17 comparative nonrandomized studies, 262 patients | OR 3.41; CI, 1.68-6.94; I2 = 0% | Very low (nonrandomized studies with increased risk of bias and imprecision due to small sample size). Owing to multiple limitations, we did not apply the large effect criteria for rating up |
| Resolution of diabetes insipidus | 11 comparative nonrandomized studies, 150 patients | OR 0.90; CI, 0.26-3.10; I2 = 0% | Very low (nonrandomized studies with increased risk of bias and imprecision due to small sample size) |
| Improvement or regression on imagining | 15 comparative nonrandomized studies, 211 patients | OR 3.13; CI, 1.54-6.36; I2 = 0% | Very low (nonrandomized studies with increased risk of bias and imprecision due to small sample size). Owing to multiple limitations, we did not apply the large effect criteria for rating up |
Abbreviation: OR, odds ratio.
When examining the effect of HD-GCs on imaging findings compared to observation, the odds of either resolution or improvement of LyHy was higher with HD-GCs (OR 3.13; 95% CI, 1.54-6.36) (Fig. 6C). Symptoms such as headache or visual symptoms were inconsistently reported according to treatment modality and therefore no comparison concerning symptomatic response to treatment could be made.
Methodological Quality and Certainty in Estimates
Of 28 comparative studies, overall methodological quality was high for selection of cohort and ascertainment of the outcome at baseline, but the comparability of the groups was inadequate for most studies (high risk of bias in selecting patients to receive a particular management strategy) (supplementary appendix 4 Table 1) (27). Ascertainment of outcomes at follow-up was not available for all identified outcomes (symptoms, pituitary function, relapse, imaging), and when assessed, around 50% of studies had an adequate methodological quality, with the rest being of intermediate or low quality, due either to inadequate ascertainment, duration, or adequacy of follow-up for a specific outcome (see supplementary appendix 4 Table 2) (27). Of the 4 noncomparative studies, methodological quality was inadequate in 3, mainly due to inadequate assessment of outcome at follow-up, duration of follow-up, or reporting (see supplementary appendix 4 Table 3) (27). Certainty in the estimates was very low for all outcomes because they were derived from nonrandomized studies and were imprecise because of small sample size (see GRADE table, supplementary appendix 3).
Discussion
In this systematic review and meta-analysis of patients with primary LyHy, we demonstrate that a greater proportion of patients had an improvement of their anterior pituitary function (partial or complete) when HD-GCs were used as first-line treatment, whereas the greatest proportion of patients with radiographic improvement or regression was noted when surgery was performed. The odds of anterior pituitary hormone recovery or radiographic improvement was 3 times greater when HD-GCs were used as first-line therapy to treat LyHy compared to observation, but patients were more likely to require additional treatment on follow-up. Notably, the proportion of patients with LyHy who recovered from DI was low and no statistically significant difference in the odds of recovery was seen between the 2 treatment modalities, HD-GCs, and observation.
LyHy has been classically reported in women and associated with pregnancy. Our results indicate a preponderance of LyHy in women (78%); however, most (80%) cases were not associated with pregnancy. The clinical relevance of autoantibody testing remains unclear as it was not systematically assessed in the studies included. Most commonly, patients presented with headache, followed by polyuria/polydipsia and hypogonadal symptoms. Our findings differ from the previous report by Caturegli et al (51), which described cases of hypophysitis from 1917 to 2016 and grouped symptoms according to underlying hormonal abnormality. In that study, patients also most commonly presented with symptoms of mass effect, but the second most common presentation was hypoadrenalism (and not symptoms of DI). The early literature on hypophysitis mainly consisted of pregnancy-related hypophysitis case reports with a minority of patients presenting with DI (52, 53). This may explain the higher frequency of polyuria/polydipsia in our study as 80% of included cases were not pregnancy related. In addition, studies assessing infundibuloneurohypophysitis alone are likely to overrepresent polyuria/polydipsia (37, 42). We found that overall, hypogonadism was the most common anterior pituitary abnormality in patients with LyHy, followed closely by central adrenal insufficiency. The reported higher proportion of adrenal insufficiency in the study by Caturegli and colleagues (51) could be also due to the inclusion of case reports and case series of fewer than 5 patients, leading to a higher selection bias, and that were excluded in our study.
LyHy is a rare disease, with the reported incidence ranging from 0.24% to 1.1% (4, 6-8, 49, 54) and a total of 1005 reported cases until 2016 (51). This makes determining optimal treatment challenging. Moreover, the definition of what constitutes treatment success (clinical, radiological, biochemical, or all 3 is not clear. Historically, surgery had been the first-line treatment, providing rapid decompression and, therefore, symptomatic relief (headache and visual disturbances) as well as pathological confirmation of the diagnosis (8, 10). Surgery is associated with a high risk of postoperative pituitary failure as the inflamed pituitary gland itself has to be removed (22). Given the invasive nature of surgery and potential complications, less invasive strategies are recommended, at least initially. Whether initial observation vs HD-GCs is preferred is not clear. We demonstrate that the use of HD-GCs had higher odds of anterior pituitary hormone recovery and improvement of magnetic resonance imaging findings compared to observation while having the lowest incidence of new anterior pituitary deficits. However, similar to surgery, HD-GCs are not free of complications with reported adverse effects in patients with LyHy including hyperglycemia, weight gain, edema, psychiatric symptoms, infection, hypokalemia, glaucoma, avascular necrosis of the hip, and osteoporosis, the majority of which are reversible on cessation of treatment (18, 22, 24, 38, 40, 45). Also, spontaneous recovery in patients using a more conservative approach was also observed (13) and therefore the risks and benefits of HD-GCs should be taken into consideration when deciding on an optimal treatment approach. With regard to the results of our meta-analysis, observation and HD-GCs both are successful initial management strategies. Initial observation might be appropriate for patients presenting with mild symptoms and or signs. HD-GCs should be used as first-line therapy in patients with severe or progressive symptoms and or signs. The main indication for surgery remains histological confirmation if the diagnosis is unclear.
Given the variable GC regimens used in the studies reviewed, it was not feasible to determine the optimal dose, duration, or type of GC to recommend. Those treated initially with HD-GCs were more likely to require additional or repeat treatment, but again this may reflect the varying dose and duration of HD-GCs. It is important to note that recurrence or progression was not explicitly defined in the studies. Alternatively, it is possible that GCs suppress the inflammatory response but do not eradicate the inciting factor. It is also possible that disease stage or chronicity may play a role in a patient’s response to HD-GCs. The mean duration of symptoms before each mode of therapy could not be determined because this was reported in only a few studies. While symptom duration before treatment is subject to recall bias, this may provide some indicator of disease chronicity. It is possible that chronic inflammation and likely scarring may not be reversible with HD-GCs compared to acute hypophysitis.
Although the majority of patients improved symptomatically, the response according to treatment modality was infrequently reported and typically not in the same manner or to the same detail as it was before treatment. Therefore, assessment of symptom outcome according to treatment could not be performed.
This is the first meta-analysis reporting outcomes according to treatment in patients with LyHy. A comprehensive literature search as well as an in-depth review of the studies included was performed. In addition, we independently selected studies and extracted data in duplicate to reduce risk for error. Furthermore, we followed the recommended approach for methodological assessment (28). Also, authors were contacted and when individual patient data were provided, we were able to exclude overlapping patients and focused on those presumed to have LyHy.
The main limitations of the study include the retrospective nature of most of the studies included, which likely contributed to the difficulties encountered regarding symptom assessment. To increase the representativeness of our population of interest, case reports and small case series (< 5 cases) were excluded. Additionally, in cases where biopsy or surgical procedure was not performed, we could not eliminate an alternative etiology of hypophysitis (eg, granulomatous hypophysitis) and had to rely on the clinical judgment and evaluation of the treating clinician. These studies were not randomized; therefore, heterogeneity in baseline characteristics of the patients receiving the 3 different treatment modalities should be taken in consideration and treatment allocation was likely subjected to selection bias. It is possible that the most severe cases of LyHy were treated more aggressively with either HD-GCs or surgery and the milder cases with a conservative approach. Consequently, the generalizability of our results to all patients with LyHy may be limited. In addition, we included studies in which patients received more than one intervention in order to compare their outcomes. However, some of these studies were not designed as a rigorous cohort study and may not have had reliable or a priori–defined numerators and denominators. Despite this, those treated with HD-GCs still saw a benefit in anterior pituitary dysfunction. A multicenter, randomized controlled trial would be necessary to overcome these limitations but may not be feasible given the rarity of the condition. Alternatively, acquisition of individual patient data through the development of rare disease registries may be required to provide more robust data as suggested by Caturegli et al (51). Publication bias is likely but we could not evaluate it statistically because of the small number of studies included per analysis and heterogeneity Last, as illustrated by the quality assessment, the comparability of subgroups was at high risk of bias in the comparative studies.
In conclusion, use of HD-GCs as first-line therapy was associated with a higher likelihood of anterior pituitary hormone recovery and regression on magnetic resonance imaging when compared to observation. However, additional treatment may be necessary on HD-GC taper or withdrawal, necessitating appropriate patient counseling regarding symptom recurrence and close clinical follow-up.
Acknowledgments
We thank all the teams who conducted the studies included in this systematic review. In particular, we would like to thank those who responded to the questions posed. Additionally, we would like to thank Larry Prokop, who performed the invaluable literature search.
Financial Support: This work was partially supported by the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health (NIH) USA (award No. K23DK121888 to I.B.). The views expressed are those of the authors and not necessarily those of the NIH USA. The sponsors had no role in designing the study; collecting, analyzing, or interpreting the data; or writing the report.
Author Contributions: Z.S., D.D., F.A., S.H.O., J.H., and D.Do. collected the data. M.H.M., I.B., and D.Do. performed the data analysis. Figures and tables were created by H.M., D.Do., and I.B. D.Do. wrote the first draft while all other authors provided critical input to data interpretation and the creation of the manuscript. The corresponding authors had full access to all the data and the final responsibility to submit for publication.
Glossary
Abbreviations
- DI
diabetes insipidus
- GCs
glucocorticoids
- HD-GCs
high-dose glucocorticoids
- IgG4
immunoglobulin G 4
- LyHy
lymphocytic hypophysitis
- OR
odds ratio
- PRISMA
Preferred Reporting Items for Systematic Reviews and Meta-Analysis
Additional Information
Disclosures: D.Do. is an investigator for and has received honoraria for advisory board participation from Corcept Therapeutics. I.B. reports advisory board participation and/or consulting with Corcept Therapeutics, Strongbridge, Sparrow Pharmaceutics, and HRA Pharma, and data and safety monitoring board participation with Adrenas Pharmaceutics outside the submitted work. The other authors have nothing to disclose.
Data Availability
Statistical code and deidentified data sets are available on request.
Clinical Trial Information: PROSPERO registration number CRD4201918989 (registered June 21, 2019).
References
- 1. Duan K, Asa SL, Winer D, Gelareh Z, Gentili F, Mete O. Xanthomatous hypophysitis is associated with ruptured Rathke’s cleft cyst. Endocr Pathol. 2017;28(1):83-90. [DOI] [PubMed] [Google Scholar]
- 2. Joshi MN, Whitelaw BC, Carroll PV. Mechanisms in endocrinology: hypophysitis: diagnosis and treatment. Eur J Endocrinol. 2018;179(3):R151-R163. [DOI] [PubMed] [Google Scholar]
- 3. Chiloiro S, Tartaglione T, Angelini F, et al. An overview of diagnosis of primary autoimmune hypophysitis in a prospective single-center experience. Neuroendocrinology. 2017;104(3):280-290. [DOI] [PubMed] [Google Scholar]
- 4. Sautner D, Saeger W, Lüdecke DK, Jansen V, Puchner MJ. Hypophysitis in surgical and autoptical specimens. Acta Neuropathol. 1995;90(6):637-644. [DOI] [PubMed] [Google Scholar]
- 5. Fehn M, Sommer C, Ludecke DK, Plockinger U, Saeger W. Lymphocytic hypophysitis: light and electron microscopic findings and correlation to clinical appearance. Endocr Pathol. 1998;9(1):71-78. [DOI] [PubMed] [Google Scholar]
- 6. Honegger J, Fahlbusch R, Bornemann A, et al. Lymphocytic and granulomatous hypophysitis: experience with nine cases. Neurosurgery. 1997;40(4):713-722. [DOI] [PubMed] [Google Scholar]
- 7. Leung GK, Lopes MB, Thorner MO, Vance ML, Laws ER Jr. Primary hypophysitis: a single-center experience in 16 cases. J Neurosurg. 2004;101(2):262-271. [DOI] [PubMed] [Google Scholar]
- 8. Caturegli P, Newschaffer C, Olivi A, Pomper MG, Burger PC, Rose NR. Autoimmune hypophysitis. Endocr Rev. 2005;26(5):599-614. [DOI] [PubMed] [Google Scholar]
- 9. De Bellis A, Bizzarro A, Bellastella A. Pituitary antibodies and lymphocytic hypophysitis. Best Pract Res Clin Endocrinol Metab. 2005;19(1):67-84. [DOI] [PubMed] [Google Scholar]
- 10. Beressi N, Beressi JP, Cohen R, Modigliani E. Lymphocytic hypophysitis. A review of 145 cases. Ann Med Interne (Paris). 1999;150(4):327-341. [PubMed] [Google Scholar]
- 11. Thodou E, Asa SL, Kontogeorgos G, Kovacs K, Horvath E, Ezzat S. Clinical case seminar: lymphocytic hypophysitis: clinicopathological findings. J Clin Endocrinol Metab. 1995;80(8):2302-2311. [DOI] [PubMed] [Google Scholar]
- 12. Howlett TA, Levy MJ, Robertson IJ. How reliably can autoimmune hypophysitis be diagnosed without pituitary biopsy. Clin Endocrinol (Oxf). 2010;73(1):18-21. [DOI] [PubMed] [Google Scholar]
- 13. Menon SK, Sarathi V, Bandgar TR, Menon PS, Goel N, Shah NS. Autoimmune hypophysitis: a single centre experience. Singapore Med J. 2009;50(11):1080-1084. [PubMed] [Google Scholar]
- 14. Khare S, Jagtap VS, Budyal SR, et al. Primary (autoimmune) hypophysitis: a single centre experience. Pituitary. 2015;18(1):16-22. [DOI] [PubMed] [Google Scholar]
- 15. Kyriacou A, Gnanalingham K, Kearney T. Lymphocytic hypophysitis: modern day management with limited role for surgery. Pituitary. 2017;20(2):241-250. [DOI] [PubMed] [Google Scholar]
- 16. Gutenberg A, Hans V, Puchner MJ, et al. Primary hypophysitis: clinical-pathological correlations. Eur J Endocrinol. 2006;155(1):101-107. [DOI] [PubMed] [Google Scholar]
- 17. Oguz SH, Soylemezoglu F, Sendur SN, et al. Clinical characteristics, management, and treatment outcomes of primary hypophysitis: a monocentric cohort. Horm Metab Res. 2020;52(4):220-227. [DOI] [PubMed] [Google Scholar]
- 18. Wang S, Wang L, Yao Y, et al. Primary lymphocytic hypophysitis: clinical characteristics and treatment of 50 cases in a single centre in China over 18 years. Clin Endocrinol (Oxf). 2017;87(2):177-184. [DOI] [PubMed] [Google Scholar]
- 19. Selch MT, DeSalles AA, Kelly DF, et al. Stereotactic radiotherapy for the treatment of lymphocytic hypophysitis. Report of two cases. J Neurosurg. 2003;99(3):591-596. [DOI] [PubMed] [Google Scholar]
- 20. Ray DK, Yen CP, Vance ML, Laws ER, Lopes B, Sheehan JP. Gamma Knife surgery for lymphocytic hypophysitis. J Neurosurg. 2010;112(1):118-121. [DOI] [PubMed] [Google Scholar]
- 21. Lupi I, Manetti L, Raffaelli V, et al. Diagnosis and treatment of autoimmune hypophysitis: a short review. J Endocrinol Invest. 2011;34(8):e245-e252. [DOI] [PubMed] [Google Scholar]
- 22. Honegger J, Buchfelder M, Schlaffer S, et al. ; Pituitary Working Group of the German Society of Endocrinology . Treatment of primary hypophysitis in Germany. J Clin Endocrinol Metab. 2015;100(9):3460-3469. [DOI] [PubMed] [Google Scholar]
- 23. Lupi I, Cosottini M, Caturegli P, et al. Diabetes insipidus is an unfavorable prognostic factor for response to glucocorticoids in patients with autoimmune hypophysitis. Eur J Endocrinol. 2017;177(2):127-135. [DOI] [PubMed] [Google Scholar]
- 24. Chiloiro S, Tartaglione T, Capoluongo ED, et al. Hypophysitis outcome and factors predicting responsiveness to glucocorticoid therapy: a prospective and double-arm study. J Clin Endocrinol Metab. 2018;103(10):3877-3889. [DOI] [PubMed] [Google Scholar]
- 25. Amereller F, Küppers AM, Schilbach K, Schopohl J, Störmann S. Clinical characteristics of primary hypophysitis—a single-centre series of 60 cases. Exp Clin Endocrinol Diabetes. 2021;129(3):234-240. [DOI] [PubMed] [Google Scholar]
- 26. Angelousi A, Cohen C, Sosa S, et al. Clinical, endocrine and imaging characteristics of patients with primary hypophysitis. Horm Metab Res. 2018;50(4):296-302. [DOI] [PubMed] [Google Scholar]
- 27. Donegan D, Saeed Z, Delivanis D, et al. Supplementary data for “Outcomes of management strategies in patients with autoimmune lymphocytic hypophysitis: a systematic review and meta-analysis.” Quality assessment appendix 1 and 2. Uploaded July 7, 2021. 10.6084/m9.figshare.14924289.v1. [DOI]
- 28. Murad MH, Sultan S, Haffar S, Bazerbachi F. Methodological quality and synthesis of case series and case reports. BMJ Evid Based Med. 2018;23(2):60-63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Wells GA, Shea B, O’Connell D, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Accessed November 28, 2021. http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp
- 30. Barendregt JJ, Doi SA, Lee YY, Norman RE, Vos T. Meta-analysis of prevalence. J Epidemiol Community Health. 2013;67(11):974-978. [DOI] [PubMed] [Google Scholar]
- 31. Murad MH. Clinical practice guidelines: a primer on development and dissemination. Mayo Clin Proc. 2017;92(3):423-433. [DOI] [PubMed] [Google Scholar]
- 32. Ahmadi J, Meyers GS, Segall HD, Sharma OP, Hinton DR. Lymphocytic adenohypophysitis: contrast-enhanced MR imaging in five cases. Radiology. 1995;195(1):30-34. [DOI] [PubMed] [Google Scholar]
- 33. Crock PA. Cytosolic autoantigens in lymphocytic hypophysitis. J Clin Endocrinol Metab. 1998;83(2):609-618. [DOI] [PubMed] [Google Scholar]
- 34. Flanagan DE, Ibrahim AE, Ellison DW, Armitage M, Gawne-Cain M, Lees PD. Inflammatory hypophysitis—the spectrum of disease. Acta Neurochir (Wien). 2002;144(1):47-56. [DOI] [PubMed] [Google Scholar]
- 35. Guo S, Wang C, Zhang J, Tian Y, Wu Q. Diagnosis and management of tumor-like hypophysitis: a retrospective case series. Oncol Lett. 2016;11(2):1315-1320. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Imber BS, Lee HS, Kunwar S, Blevins LS, Aghi MK. Hypophysitis: a single-center case series. Pituitary. 2015;18(5):630-641. [DOI] [PubMed] [Google Scholar]
- 37. Imura H, Nakao K, Shimatsu A, et al. Lymphocytic infundibuloneurohypophysitis as a cause of central diabetes insipidus. N Engl J Med. 1993;329(10):683-689. [DOI] [PubMed] [Google Scholar]
- 38. Korkmaz OP, Sahin S, Ozkaya HM, et al. Primary hypophysitis: experience of a single tertiary center. Exp Clin Endocrinol Diabetes. 2021;129(1):14-21. [DOI] [PubMed] [Google Scholar]
- 39. Panigrahi M, Kumari M, Vooturi S. Corticosteroids in the management of lymphocytic hypophysitis: case series. Neurol India. 2018;66(2):400-404. [DOI] [PubMed] [Google Scholar]
- 40. Park SM, Bae JC, Joung JY, et al. Clinical characteristics, management, and outcome of 22 cases of primary hypophysitis. Endocrinol Metab (Seoul). 2014;29(4):470-478. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Pérez G, Almeda-Valdés P, Cuevas-Ramos D, Juárez-Comboni SC, Higuera-Calleja J, Gómez-Pérez FJ. Hypophysitis autoimmune. Case series and literature review [article in Spanish]. Gac Med Mex. 2013;149(2):229-236. [PubMed] [Google Scholar]
- 42. Schaefers J, Cools M, De Waele K, et al. Clinical presentation and outcome of children with central diabetes insipidus associated with a self-limited or transient pituitary stalk thickening, diagnosed as infundibuloneurohypophysitis. Clin Endocrinol (Oxf). 2017;87(2):171-176. [DOI] [PubMed] [Google Scholar]
- 43. Tauziede-Espariat A, Polivka M, Bouazza S, et al. The prevalence of IgG4-positive plasma cells in hypophysitis: a possible relationship to IgG4-related disease. Clin Neuropathol. 2015;34(4):181-192. [DOI] [PubMed] [Google Scholar]
- 44. Tirosh A, Hirsch D, Robenshtok E, et al. Variations in clinical and imaging findings by time of diagnosis in females with hypopituitarism attributed to lymphocytic hypophysitis. Endocr Pract. 2016;22(4):447-453. [DOI] [PubMed] [Google Scholar]
- 45. Zhu Q, Qian K, Jia G, et al. Clinical features, magnetic resonance imaging, and treatment experience of 20 patients with lymphocytic hypophysitis in a single center. World Neurosurg. 2019;127:e22-e29. [DOI] [PubMed] [Google Scholar]
- 46. Heaney AP, Sumerel B, Rajalingam R, Bergsneider M, Yong WH, Liau LM. HLA markers DQ8 and DR53 are associated with lymphocytic hypophysitis and may aid in differential diagnosis. J Clin Endocrinol Metab. 2015;100(11):4092-4097. [DOI] [PubMed] [Google Scholar]
- 47. Iuliano SL, Laws ER. The diagnosis and management of lymphocytic hypophysitis. Expert Rev Endocrinol Metab. 2011;6(6):777-783. [DOI] [PubMed] [Google Scholar]
- 48. Kristof RA, Van Roost D, Klingmüller D, Springer W, Schramm J. Lymphocytic hypophysitis: non-invasive diagnosis and treatment by high dose methylprednisolone pulse therapy? J Neurol Neurosurg Psychiatry. 1999;67(3):398-402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Nakata Y, Sato N, Masumoto T, et al. Parasellar T2 dark sign on MR imaging in patients with lymphocytic hypophysitis. AJNR Am J Neuroradiol. 2010;31(10):1944-1950. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50. Honegger J, Schlaffer S, Menzel C, et al. ; Pituitary Working Group of the German Society of Endocrinology . Diagnosis of primary hypophysitis in Germany. J Clin Endocrinol Metab. 2015;100(10):3841-3849. [DOI] [PubMed] [Google Scholar]
- 51. Caturegli P, Di Dalmazi G, Lombardi M, et al. Hypophysitis secondary to cytotoxic T-lymphocyte-associated protein 4 blockade: insights into pathogenesis from an autopsy series. Am J Pathol. 2016;186(12):3225-3235. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52. Cosman F, Post KD, Holub DA, Wardlaw SL. Lymphocytic hypophysitis. Report of 3 new cases and review of the literature. Medicine (Baltimore). 1989;68(4):240-256. [DOI] [PubMed] [Google Scholar]
- 53. Honegger J, Giese S. Acute pituitary disease in pregnancy: how to handle hypophysitis and Sheehan’s syndrome. Minerva Endocrinol. 2018;43(4):465-475. [DOI] [PubMed] [Google Scholar]
- 54. Buxton N, Robertson I. Lymphocytic and granulocytic hypophysitis: a single centre experience. Br J Neurosurg. 2001;15(3):242-245; discussion 245-246. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Statistical code and deidentified data sets are available on request.
Clinical Trial Information: PROSPERO registration number CRD4201918989 (registered June 21, 2019).