Predictors of Quality of Life in 165 Patients With Acromegaly

Results From a Single-Center Study

Ilonka Kreitschmann-Andermahr, MD; Michael Buchfelder, MD; Bernadette Kleist, CN; Johannes Kohlmann, MD; Christa Menzel, RN; Rolf Buslei, MD; Maria Kołtowska-Häggström, MD; Christian Strasburger, MD; Sonja Siegel, CP


Endocr Pract. 2017;23(1):79-88. 

In This Article


A total of 417 patients with acromegaly caused by a GH-secreting pituitary adenoma, operated upon between 2000 and 2012 at the Department of Neurosurgery, University of Erlangen-Nuremberg, were identified from patient records. Inclusion criteria were: age between 18 and 70 years at the time of recruitment, sufficient command of the German language to complete health-related questionnaires, and absence of an active psychotic disorder. Of those 417 patients, 103 did not meet the inclusion criteria. The remaining 314 received an explanatory letter as to the purpose of the study, a consent form, and a set of questionnaires. A total of 177 patients returned the questionnaires, of which 12 were excluded for formal reasons, such as largely incomplete data (n = 2), insufficient command of the German language (n = 1), missing consent forms (n = 4), being treated for a long time in a country not compatible with German healthcare standards (n = 1), or failure to meet inclusion criteria at second check (n = 4). The 165 patients included in the analysis corresponded to a 52.5% inclusion rate.

The 165 included patients answered a self-developed survey, containing a set of 54 questions designed to gain detailed and standardized information about each participant's age, gender, height, weight, level of education, occupational status (including shift work), alcohol and tobacco consumption, medical history, and interaction with healthcare. Specifically, the number of doctors visited, their respective specialty fields, reasons for consultation, timing of onset of symptoms, date at which medical assistance was first sought, diagnosis, types of treatment, and follow-up treatment(s) were explored. The patient-reported outcome questionnaire was designed to capture the diagnostic and therapeutic process of German patients with acromegaly from the patient's perspective. The diagnostic and therapeutic pathways of these patients have recently been published.[11]

For the current investigation, the following standardized, self-reported questionnaires capturing QoL, depression, and embitterment were analyzed in conjunction with the clinical information: the Short-Form Health Survey (SF-36),[12] the revised Beck Depression Inventory (BDI-II),[13] and the Bern Embitterment Inventory (BEI).[14] The German-validated versions of the SF-36[15] and BDI-II[16,17] were used; this was not necessary for the BEI, as this questionnaire was originally developed in German.

The SF-36 is a frequently employed, well-established generic QoL questionnaire. It consists of 36 questions, which measure eight domains: vitality, physical functioning, bodily pain, general health perceptions, physical role functioning, emotional role functioning, social role functioning, and mental health. It differentiates between physical and mental QoL by enabling the calculation of two separate scores: the Physical Component Summary (SF-36 PCS) score, and the Mental Component Summary (SF-36 MCS) score. The raw values were transformed into standardized scores to allow for comparison with the age- and sex-matched reference values of the German normative population.[18] The transformed subscales are scaled from 0 to 100; the transformed summary scores have a mean of 50 and SD of 10. Higher scores indicate a better QoL.[15] The SD score (SDS) was also recorded. A score > −2 SDS below the mean indicates a severely impaired QoL, and a score between −1 and −2 SDS indicated an impaired QoL. The BDI-II is a self-reported 21-item measure assessing depression. Respondents rated perceived severity of depressive symptoms using a 4-point Likert scale (0 to 3, with 0 indicating absence of a depressive symptom and 3 indicating a severe depressive symptom). A higher score indicates more severe depressive symptoms. There are cutoffs for no (score 0 to 8), minor (score 9 to 13), mild (score 14 to 19), moderate (score 20 to 28), and severe (score 29 to 63) depressive symptoms.[13]

The BEI[14] measures embitterment in four dimensions: emotional embitterment, performance-related embitterment, pessimism/hopelessness, and misanthropy, which together yield a total BEI score. Higher values indicate a higher degree of embitterment. The developers of the tool defined, based on the reference values of healthy subjects, four degrees of embitterment, with raw scores from 0.00 to 0.99 indicating a low, 1.00 to 1.99 a medium, 2.00 to 2.99 an above-average, and 3.00 to 3.99 an extreme degree of embitterment.

The study followed the Declaration of Helsinki and was approved by the Friedrich-Alexander University Erlangen-Nuremberg Ethics Committee (Re. no. 144_12 B). All participating patients provided written informed consent.

Statistical Analyses

Statistical analyses were carried out using SPSS 22.0 (Statistical Package for the Social Sciences, SPSS Inc, Armonk, NY) for Windows. Interval scaled data were described by way of means and SD, and categorical data by relative frequency given in percentage. The standardized SF-36 scores and the raw scores of other questionnaires were used for analysis. Missing variables were regarded as missing at random. The assumptions of linearity, normality, homoscedasticity, independence of errors, and absence of multicollinearity were checked. Although minor violations were observed, the main criteria for calculating the regression models were met.

In the first step of the data analysis, we explored the association between clinical variables such as age, BMI, current insulin-like growth factor 1 (IGF-1) level, and previous radiotherapy with QoL, depression, and embitterment. To this effect, the responses to the general questions were correlated (Pearson's r) with the SF-36 PCS score, SF-36 MCS score, BDI-II score, and BEI score as criterion variables (i.e., the variables that are later predicted in the regression model). In the second step, those clinical variables that were significantly correlated with the criterion variables were used as predictors in the regression analysis. A separate stepwise regression model for each of the four criterion variables was developed. Therefore, the variable with the highest correlation to the criterion variable was entered first. In subsequent regression steps, only variables with a significant ΔR2 , which specifies the amount of explained variance that is added by the entry of the respective variable, were entered. As a measure of the degree to which the final regression model predicts the criterion variables, the overall amount of explained variance (R2 ) is presented. For each regression model, regression coefficients (β) and standardized regression coefficients (BETA), which indicate whether a predictor is positively or negatively related to the criterion, are reported. A level of P≤.05 was considered to be significant.