Feature Request: Calculate Creatinine Clearance

[jeffreyad]

Feature Idea

This function would calculate Creatinine Clearance (CRCL). This is used renal function assessment. It could be used to create a parameter for ADLB or a baseline value. The baseline values are typically used in PK programming for NCA or POPPK. It might be similar to the functions in derive_advs_params.R such as derive_param_bmi() and compute_bmi()

This is the formula I have typically seen used:
For CREATBL in umol/L:
For SEX = 'F': then CRCL = ((140 - AAGE) * WTBL * 1.04) / CREATBL
For SEX = 'M': then CRCL = ((140 - AAGE) * WTBL * 1.23) / CREATBL

There could be the option to use different formulas if desired.

It might also be worth calculating EGFR

Relevant Input

base <- tribble(
  ~STUDYID, ~USUBJID, ~AGE, ~SEX, ~RACE, ~WTBL, ~CREATBL, ~CREATBLU,
  "P01", "P01-1001", 45, "M", "WHITE", 85, 90, "umol/L",
  "P01", "P01-1002", 52, "F", "BLACK OR AFRICAN AMERICAN", 68, 70, "umol/L",
  "P01", "P01-1003", 67, "M", "BLACK OR AFRICAN AMERICAN", 85, 77, "umol/L",
  "P01", "P01-1004", 76, "F", "ASIAN", 60, 65, "umol/L",
)
# A tibble: 4 × 8
  STUDYID USUBJID    AGE SEX   RACE                       WTBL CREATBL CREATBLU
  <chr>   <chr>    <dbl> <chr> <chr>                     <dbl>   <dbl> <chr>   
1 P01     P01-1001    45 M     WHITE                        85      90 umol/L  
2 P01     P01-1002    52 F     BLACK OR AFRICAN AMERICAN    68      70 umol/L  
3 P01     P01-1003    67 M     BLACK OR AFRICAN AMERICAN    85      77 umol/L  
4 P01     P01-1004    76 F     ASIAN                        60      65 umol/L  

Relevant Output

# A tibble: 4 × 10
  STUDYID USUBJID    AGE SEX   RACE                       WTBL CREATBL CREATBLU CRCLBL CRCLBLU
  <chr>   <chr>    <dbl> <chr> <chr>                     <dbl>   <dbl> <chr>     <dbl> <chr>  
1 P01     P01-1001    45 M     WHITE                        85      90 umol/L    110.  mL/min 
2 P01     P01-1002    52 F     BLACK OR AFRICAN AMERICAN    68      70 umol/L     88.9 mL/min 
3 P01     P01-1003    67 M     BLACK OR AFRICAN AMERICAN    85      77 umol/L     99.1 mL/min 
4 P01     P01-1004    76 F     ASIAN                        60      65 umol/L     61.4 mL/min 

Reproducible Example/Pseudo Code

crcl <- base %>%
  mutate(
    CRCLBL = case_when(
      SEX == "F" ~ ((140 - AGE) * WTBL * 1.04) / CREATBL,
      SEX == "M" ~ ((140 - AGE) * WTBL * 1.23) / CREATBL
    ),
    CRCLBLU = "mL/min",
  )

[airdjan]

Based on the data set above, I have added the calculation of the GFR value:

CKD-EPI Creatinine Equation (2021)

$$eGFR = 142 \times min(S_{cr}/\kappa, 1)^\alpha \times max(S_cr/\kappa, 1)^{-1.200} \times 0.9938^{Age} \times 1.012 [if female]$$$

where:
$S_{cr}$ = standardized serum creatinine in mg/dL
$\kappa$ = 0.7 (females) or 0.9 (males)
$\alpha$ = -0.241 (female) or -0.302 (male)
min($S_{cr}/\kappa$, 1) is the minimum of $S_{cr}/\kappa$ or 1.0
max($S_{cr}/\kappa$, 1) is the maximum of $S_{cr}/\kappa$ or 1.0
Age (years)

Source = https://www.kidney.org/content/ckd-epi-creatinine-equation-2021

Relevant Input

base <- tribble(
  ~STUDYID, ~USUBJID, ~AGE, ~SEX, ~RACE, ~WTBL, ~CREATBL, ~CREATBLU, ~SCR,
  "P01", "P01-1001", 45, "M", "WHITE", 85, 90, "umol/L", 1.2,
  "P01", "P01-1002", 52, "F", "BLACK OR AFRICAN AMERICAN", 68, 70, "umol/L", 1.3,
  "P01", "P01-1003", 67, "M", "BLACK OR AFRICAN AMERICAN", 85, 77, "umol/L", 1.1,
  "P01", "P01-1004", 76, "F", "ASIAN", 60, 65, "umol/L", 1.2)

# A tibble: 4 × 9
  STUDYID USUBJID    AGE SEX   RACE                       WTBL CREATBL CREATBLU   SCR
  <chr>   <chr>    <dbl> <chr> <chr>                     <dbl>   <dbl> <chr>    <dbl>
1 P01     P01-1001    45 M     WHITE                        85      90 umol/L     1.2
2 P01     P01-1002    52 F     BLACK OR AFRICAN AMERICAN    68      70 umol/L     1.3
3 P01     P01-1003    67 M     BLACK OR AFRICAN AMERICAN    85      77 umol/L     1.1
4 P01     P01-1004    76 F     ASIAN                        60      65 umol/L     1.2

Relevant Output

# A tibble: 4 × 11
# Rowwise: 
  STUDYID USUBJID    AGE SEX   RACE                       WTBL CREATBL CREATBLU   SCR  EGFR EGFRI        
  <chr>   <chr>    <dbl> <chr> <chr>                     <dbl>   <dbl> <chr>    <dbl> <dbl> <chr>        
1 P01     P01-1001    45 M     WHITE                        85      90 umol/L     1.2  76.0 mL/min/1.73m2
2 P01     P01-1002    52 F     BLACK OR AFRICAN AMERICAN    68      70 umol/L     1.3  49.5 mL/min/1.73m2
3 P01     P01-1003    67 M     BLACK OR AFRICAN AMERICAN    85      77 umol/L     1.1  73.6 mL/min/1.73m2
4 P01     P01-1004    76 F     ASIAN                        60      65 umol/L     1.2  46.9 mL/min/1.73m2

Reproducible Example/Pseudo Code

calculate_CKD_EPI <- function(age, sex, scr) {
  if (is.na(age) || is.na(sex) || is.na(scr) ||
      !is.numeric(age) || !is.numeric(scr)) {
    return(NA_real_)
  }
  
  kappa <- case_when(grepl("F",sex) ~ 0.7,
                     grepl("M",sex) ~ 0.9,
                     TRUE ~ NA_real_)
  
  alpha <- case_when(grepl("F",sex) ~  -0.241,
                     grepl("M",sex) ~  -0.302,
                     TRUE ~ NA_real_)
  
  gender_coefficent <- case_when(grepl("F",sex) ~ 1.012,
                                TRUE ~ 1)
  
  gfr <- 142 * min(scr/kappa, 1)^(alpha) * 
    max(scr/kappa, 1)^(-1.200) * 
    0.9938^age * 
    gender_coefficent
  
  return(gfr)
}

#additional compute of the CKD Stages if needed
calculate_CKD_stage <- function(egfr) {
  ckd_stage <- case_when(egfr >= 90 ~ "Stage 1", 
                         (egfr <= 89) | (egfr >= 60) ~ "Stage 2", 
                         (egfr <= 59) | (egfr >= 30) ~ "Stage 3", 
                         (egfr <= 29) | (egfr >= 15) ~ "Stage 4", 
                         egfr < 15 ~ "Stage 5",
                         TRUE ~ "No Stage")
  return(ckd_stage)
}

base %>% 
  rowwise() %>% 
  mutate(EGFR = calculate_CKD_EPI(AGE,SEX,SCR),
         EGFRI = "mL/min/1.73m2")

[bms63]

This is awesome @airdjan ! Thank you for putting together. Let's discuss at tomorrow's PK Working group meeting.

[jeffreyad]

@airdjan @bms63 I added an initial compute_crcl() which we can discuss in the PK meeting. We could potentially add the CKD-EPI equation to the same function or a separate function.