Database
Real-world data from MID-NET®, a reliable and valuable database in Japan [14, 15], were used for analysis in this study because MID-NET® stores electronic medical records, administrative claim data, and diagnosis procedure combination (DPC) data for over 6.05 million patients (as of December 2022) in cooperation with 10 healthcare organizations, including 23 university hospitals and regional core hospitals. The study period spanned from January 1, 2009, to March 31, 2019.
Utilizing MID-NET® for this study was approved on February 19, 2020, through a discussion by the expert committee of MID-NET® [16] and the actual data extraction from MID-NET® for analysis was carried out in the week of May 22, 2020.
Cohort and study design
In this study, a new-user cohort design was selected for considering the degree of renal dysfunction and abnormal serum calcium (Ca) levels after BP administration. Specifically, as shown in Fig. 1, patients who met all of the following conditions were included in this study; 1) patients prescribed BPs targeted in this study during the study period, 2) patients with a record of an osteoporosis-related diagnosis in the same month as t0 (the earliest prescription date of BPs), 3) patients with initial medical records at least 90 days before t0, and 4) patients with a record of serum creatinine (Cr) or estimated glomerular filtration rate (eGFR) from 90 days before t0 to the day before t0 (baseline period). For the analysis, patients who met one or more of the following criteria were excluded to select an appropriate population without a higher risk of hypocalcemia; 1) patients prescribed multiple BPs at t0, 2) patients with a record of corrected serum Ca level < 8.00 mg/dL during the baseline period, 3) patients with a record of an episode of primary hyperparathyroidism during the baseline period, 4) patients prescribed denosumab (genetical recombination) at least once during and after the baseline period, 5) patients with a record of an episode of acute pancreatitis or sepsis during and after the baseline period, and 6) patients prescribed at least one dose of asfotase alfa (genetical recombination), cinacalcet hydrochloride, evocalcet, or etelcalcetide hydrochloride during and after the baseline period (see Additional file 1 for more details of this study design).
Flow chart for patient selection
The target BPs in this study included all BPs (not only oral preparations but also intravenous preparations) with an indication for osteoporosis marketed in Japan during the study period; i.e., alendronate sodium hydrate, etidronate disodium, ibandronate sodium hydrate, minodronic acid hydrate, sodium risedronate hydrate, and zoledronic acid hydrate.
Definition of exposure
In order to evaluate the relationship between the incidence of hypocalcemia and decreased renal function, renal function was stratified into four categories based on eGFR value: normal (eGFR ≥ 90 mL/min/1.73m2), mild (60 mL/min/1.73m2 ≤ eGFR < 90 mL/min/1.73m2), moderate (30 mL/min/1.73m2 ≤ eGFR < 60 mL/min/1.73m2), severe (eGFR < 30 mL/min/1.73m2). In cases where eGFR values were not recorded, the values were calculated from serum Cr values using the following formula widely used in Japan [17, 18]: 194 × serum creatinine−1.094 × age−0.287 (× 0.739 for women) (mL/min/1.73m2). In instances where multiple eGFR values were available on the same day, the mean eGFR value was used. The baseline eGFR value was determined as the mean of the two eGFR values closest to t0 during the baseline period. If only one examination was performed during the baseline period, that single value was used as the baseline eGFR. No data imputations for missing values were made in calculating eGFR.
Definition of prescription and follow-up periods
The prescription period encompassed the start date (t0) and duration of the prescription with a gap and a grace period, in consideration of prescription interval for each BP and a deviation from the scheduled visit time, etc. The gap and the grace periods were equally set based on the information in the PI; i.e., 14 days for BPs prescribed either once a week or once a day, 56 days for BPs prescribed once every 4 weeks, 60 days for BPs prescribed once a month, and 90 days for BPs prescribed once a year. Consequently, two prescriptions for the same drug were considered continuous if the latter prescription date fell within the gap period of the former prescription date.
The follow-up period commenced at t0 and concluded at the earliest date of the following; 1) the end date of the prescription period, 2) the day before the start date of another different BP prescription from t0, 3) the date of changing renal function category defined as the second date of two consecutive changes of a different category from the baseline, or 4) the end date of the study period (March 31, 2019).
Definition of outcome
“Hypocalcemia” was defined as a corrected serum Ca level < 8.00 mg/dL, as per Payne’s equation, a widely employed method in Japan [19, 20]. If multiple values were recorded on the same day, the minimum serum Ca value and the maximum albumin (Alb) value were used. To determine the corrected serum calcium level, the serum Alb value measured within 14 days of the serum Ca measurement was selected, with preference given to the value closest to the date of serum Ca measurement. No data imputations for missing values were made in calculating serum Ca value.
Definition of covariates
The covariates used in this study included gender (male/female), age (age < 65 years, 65 years to < 75 years, 75 years and older), and complications (hypoparathyroidism, vitamin D deficiency, magnesium disorders) as well as concomitant drugs (elcatonin, steroids excluding topical preparations, calcium preparations excluding topical preparations, vitamin D preparations excluding topical preparations, sorafenib tosilate, lenvatinib mesilate, vandetanib, enviomycin sulfate, and monobasic sodium phosphate monohydrate/dibasic sodium phosphate anhydrous) among diseases and drugs known to be associated with the risk of hypocalcemia [8, 21, 22]. Data at t0 for sex and age, and at the baseline period for complications and concomitant drugs were used for analysis.
Statistical analysis
Patient background data, including covariates, each active ingredient of BPs at t0, concomitant drugs known to cause osteoporosis during the follow-up period, and the calendar year of t0 were tabulated.
The incidence rate of hypocalcemia (/patient-year) in each group, the crude hazard ratio (cHR) and adjusted hazard ratio (aHR; with adjustment for the covariates described above) of each group to the normal group (Cox proportional hazards model) were calculated. These analyses were also performed for each active ingredient of BPs at t0.
In addition, sensitivity analysis was conducted by changing the outcome definition from “corrected serum Ca level < 8.00 mg/dL” in the primary analysis to “20% or more reduction of corrected serum Ca level from the baseline.” The baseline serum Ca level was defined as the value closest to t0 among the corrected serum Ca values in the baseline period. Furthermore, to check the impacts of the lack of laboratory tests during the BP prescription period, the same analysis as the primary was conducted only in patients with more than one laboratory test result per 30 days during the follow-up period. In this population, the median (first quartile, third quartile) frequency for laboratory tests was 1.7 (1.4, 3.2) for the normal, 1.5 (1.4, 3.2) for the mild, 1.8 (1.4, 3.8) for the moderate, and 2.2 (1.4, 5.6) for the severe groups. The aHRs were also calculated when dialysis patients who had a record of dialysis before t0 were excluded from the cohort of the primary analysis.
SAS version 9.4 (SAS Institute, Cary, NC, USA) was used for all analysis.
- The Renal Warrior Project. Join Now
- Source: https://bmcnephrol.biomedcentral.com/articles/10.1186/s12882-024-03553-7