'Femi Fajimi

From “Raw TFLs” to Regulator-Ready Narratives: A Medical Writer’s Walkthrough (Antifungal Case Study) 

By Femi Fajimi | 10 September 2025 

(Educational content. Drug, data, and outputs below are fictional but formatted to mirror real deliverables.) 

Introduction 

Sometimes, when I am asked to support some freelance writers with their ongoing projects, particularly when it involves reviewing and interpreting TFLs (Tables, Figures, and Listings), these tasks can be intense, requiring careful attention, tightening the belt, and going line by line to proofread, verify, and ensure every number aligns with the protocol and SAP. 

When a sponsor hands over TFLs, your job as a regulatory writer is to transform those statistical outputs into clear, neutral, and fully traceable narratives for the Clinical Study Report (CSR; ICH E3) and the CTD summaries (ICH M4E(R2) Modules 2.7.3/2.7.4). That means: 

• Understanding what each TFL represents and how it was derived (as per the SAP). 

• Checking analysis sets, denominators, visit windows, and intercurrent-event handling (ICH E9/E9(R1)). 

• Writing factual interpretations that align line-by-line with TFLs, without overclaiming. 

Below is a comprehensive “starter pack” of TFLs for a fictional Phase 3 antifungal, accompanied by step-by-step commentary on what to look for and how to translate the outputs into submission-ready text. 

Study overview (fictional) 

Study ID: FGN-301 

Investigational product: Fungimycin (FGN-201), novel antifungal (IV → optional oral step-down) 

Comparator: Caspofungin (active control) 

Indication: Candidaemia/invasive candidiasis (adults) 

Design: Phase 3, randomised, double-blind, double-dummy, non-inferiority 

Primary endpoint: All-cause mortality at Day 30 (mITT) 

Key secondary endpoints: Global response at End of Therapy (EOT); time to blood-culture clearance 

Safety focus: TEAEs/SAEs, Grade ≥3 AEs, hepatic labs (Hy’s-Law assessment), ECG/QT 

Total randomised: N=520 (Fungimycin 259; Caspofungin 261) 

Standards followed (for the training example): ICH E3 (CSR structure), ICH M4E(R2) (CTD efficacy), ICH E9 + E9(R1) (statistical principles & estimands), MedDRA coding for AEs, CTCAE grading conventions, and EORTC/MSGERC criteria for invasive fungal disease case definitions. 

What typically arrives in your inbox (handover pack) 

Final TFLs (labelled, versioned, page-numbered) 

SAP (+ amendments) and Protocol (+ amendments) 

Derivation/specification documents (e.g., how global response was defined; how “missing vital status” handled) 

Randomisation & stratification details 

Programming notes (e.g., imputation rules, censoring definitions) 

Data review memos (e.g., decisions on protocol deviations) 

Tip: Before writing, create a Traceability Matrix: Endpoint → SAP section → TFL ID(s) → CSR section → CTD section → Listing(s) supporting anomalies. It will save you during QC and inspections. 

The TFLs (starter set, realistically detailed) 

Important: All numbers below are fictional but formatted to real sponsor outputs. Analysis sets, rules, and footnotes are written in a regulatory style for training purposes. 

Table 5.1 — Subject Disposition (All Randomised) 

Category	Fungimycin (N=259)	Caspofungin (N=261)	Total (N=520)
Randomised	259	261	520
Received ≥1 dose (Safety set)	258 (99.6%)	260 (99.6%)	518 (99.6%)
mITT (randomised + culture-confirmed + ≥1 dose)	259 (100%)	261 (100%)	520 (100%)
Completed study	222 (85.7%)	220 (84.3%)	442 (85.0%)
Discontinued study early	37 (14.3%)	41 (15.7%)	78 (15.0%)
– Adverse event	15 (5.8%)	12 (4.6%)	27 (5.2%)
– Lack of efficacy	10 (3.9%)	13 (5.0%)	23 (4.4%)
– Withdrawal of consent	6 (2.3%)	9 (3.4%)	15 (2.9%)
– Other	6 (2.3%)	7 (2.7%)	13 (2.5%)

Footnotes: 

• mITT = all randomised patients with culture-confirmed Candida at baseline who received ≥1 dose. 

• Completed study = completed Day-42 visit or died earlier, per Protocol §6.2. 

• Early discontinuation reasons classified per eCRF categories. 

Explanation and interpretation 

This table shows how patients moved through the trial, from randomisation to treatment exposure, analysis set inclusion, completion, and reasons for early discontinuation. It is one of the most critical early TFLs in a CSR (ICH E3 §11.1), as it confirms patient accountability. 

1. Randomised population: 520 patients were randomly assigned, evenly split between the two arms (259 vs. 261). This balance supports comparability of treatment groups. 

2. Safety set: Nearly all patients received at least one dose (99.6% both arms), ensuring that safety analyses capture the whole randomised population with minimal attrition bias. 

3. mITT set: Defined here as randomised + culture-confirmed + ≥1 dose, and notably all patients qualified (100%), meaning no exclusions for missing baseline cultures. This simplifies efficacy interpretation. 

4. Study completion: Around 85% completed the study (85.7% Fungimycin vs 84.3% caspofungin). 

5. Early discontinuation: Rates were similar (14.3% vs 15.7%). 

• Adverse events: Slightly higher in Fungimycin (5.8% vs 4.6%), though not a large imbalance. 

• Lack of efficacy: Comparable across groups (3.9% vs 5.0%). 

• Consent withdrawal/other: Low and consistent. 

Overall, there are no concerning imbalances, and discontinuation patterns appear balanced across treatment arms. 

Regulatory writer’s role 

• Traceability: Ensure denominators are consistent with SAP-defined analysis sets. 

• Clarity in CSR: For example: “Of 520 randomised patients (259 Fungimycin, 261 caspofungin), 518 (99.6%) received at least one dose and 100% qualified for the mITT set. Study completion rates were similar between arms (85.7% vs 84.3%). Early discontinuation occurred in 14–16% of patients, most commonly due to adverse events (5.8% vs 4.6%) or lack of efficacy (3.9% vs 5.0%).” 

• Forward linkage: Flag that AE-related discontinuations will be discussed in the Safety Results and lack-of-efficacy discontinuations are relevant to interpreting primary endpoint robustness. 

Table 6.1 — Baseline Demographics & Disease Characteristics (All Randomised Patients) 

Characteristic	Fungimycin (N=259)	Caspofungin (N=261)
Age, years, mean (SD)	58.2 (15.1)	58.5 (14.8)
Female, n (%)	107 (41.3)	108 (41.4)
ICU at baseline, n (%)	80 (30.9)	78 (29.9)
APACHE II, mean (SD)	14.6 (6.1)	14.9 (6.0)
Neutropenic, n (%)	47 (18.1)	45 (17.2)
Infection type, n (%) candidaemia only / deep-seated	198 (76.4) / 61 (23.6)	201 (77.0) / 60 (23.0)
Baseline species, n (%) C. albicans / C. glabrata / C. parapsilosis / C. tropicalis / C. auris / Other	112 / 62 / 34 / 21 / 6 / 24	110 / 64 / 36 / 20 / 6 / 25

Footnotes: 

• APACHE II scores were calculated centrally using the protocol-defined algorithm. 

• Species identification was based on central laboratory cultures at screening. 

• Deep-seated infections included intra-abdominal, hepatosplenic, and other tissue-invasive presentations. 

Explanation and interpretation 

This table presents demographic and baseline disease characteristics of the randomised population, which help establish comparability of treatment arms before outcomes are assessed. 

• Age and sex: Groups were well balanced, with mean age around 58 years and ~41% female in both arms. 

• ICU at baseline: About 31% overall, similar across arms, indicating a critically ill subset was equally distributed. 

• APACHE II scores: Means around 15 in both groups, suggesting comparable illness severity at baseline. 

• Neutropenia: Present in ~18% of patients in each arm, balanced across groups. 

• Infection type: Around three-quarters had candidaemia only, and one-quarter had deep-seated infection; distribution was nearly identical between groups. 

• Baseline Candida species: The most common was C. albicans, followed by C. glabrata and C. parapsilosis. Distribution of species was nearly identical across arms, including rare C. auris cases (6 in each group). 

Overall, no clinically relevant imbalances were observed. The randomisation process appears to have achieved good prognostic balance, supporting the validity of comparative analyses. 

Regulatory writer’s role 

• Traceability: Verify definitions (e.g. APACHE II algorithm, neutropenia criteria) against the SAP and protocol. 

• Clarity in CSR: For example, “Baseline characteristics were comparable between groups. Mean age was 58 years, with ~41% female. Approximately 31% were admitted to the ICU at baseline, and the mean APACHE II score was ~15 in both groups, indicating similar severity of illness. Infection type and baseline Candida species distribution were balanced across arms, with C. albicans most frequent. No imbalances likely to confound treatment comparisons were identified.” 

• Forward linkage: Note that a balanced species distribution strengthens the validity of pooled efficacy analyses, while the balance of ICU/illness severity is important for interpreting mortality outcomes. 

Table 14.1.1 — Primary Efficacy: Day-30 All-Cause Mortality (mITT) 

Endpoint	Fungimycin (N=259)	Caspofungin (N=261)	Effect
Deaths by Day 30, n (%)	18 (6.9)	22 (8.4)	Risk difference −1.5% (95% CI −5.5, +2.5)
Kaplan–Meier at Day 30, %	7.0	8.3	—
Time-to-death HR (95% CI)*	—	0.83 (0.45, 1.51)	—

Footnotes: 

• Analysis set: mITT (randomised + culture-confirmed + ≥1 dose). 

• Non-inferiority (NI) margin: 10% (absolute). Alpha: one-sided 0.025. 

• Cox model stratification: neutropenia (yes/no), baseline species group (C. auris vs. others), and region; reference arm: caspofungin. 

• Missing data: Patients with unknown vital status at Day 30 were counted as deaths (conservative approach, SAP §8.2.1). 

• Intercurrent events: Rescue antifungal use did not alter mortality classification (treatment-policy strategy). 

Explanation and interpretation 

This table presents the primary endpoint, all-cause mortality at Day 30, in the modified intention-to-treat (mITT) population. 

• Event rates: 6.9% (18/259) of Fungimycin patients and 8.4% (22/261) of caspofungin patients died by Day 30. 

• Risk difference: −1.5% (95% CI:- 5.5%, +2.5 %). The upper CI limit (+2.5%) is well below the prespecified NI margin of 10%, demonstrating non-inferiority of Fungimycin. 

• Kaplan–Meier estimate: Mortality probabilities were consistent with crude rates (7.0% vs 8.3%). 

• Time-to-event analysis: Hazard ratio 0.83 (95% CI 0.45, 1.51) suggests no excess risk with Fungimycin, though CI crosses 1.0. 

Overall, the data show that Fungimycin met the primary non-inferiority objective with respect to Day-30 mortality. 

Writer’s checks 

• Confirm NI margin (10%) and analysis set (mITT) match protocol and SAP definitions. 

• Ensure CI upper bound (+2.5%) is interpreted correctly: non-inferiority met, but not superiority. 

• Verify that the conservative handling of missing vital status is applied consistently. 

• Verify that the treatment policy strategy (rescue = on-treatment) is reported clearly. 

• Cross-check subgroup analyses (e.g., neutropenic vs non-neutropenic) show no contradictory patterns. 

CSR example text 

“In the mITT population, all-cause mortality by Day 30 occurred in 6.9% of Fungimycin patients and 8.4% of caspofungin patients. The risk difference was −1.5% (95% CI −5.5, +2.5), with the upper confidence bound below the prespecified 10% non-inferiority margin, thereby demonstrating non-inferiority of Fungimycin. Kaplan–Meier estimates and time-to-event analyses were consistent with these findings.” 

Table 14.2.1 — Key Secondary Endpoint: Global Response at End of Therapy (mITT Population) 

Category	Fungimycin (N=259)	Caspofungin (N=261)
Success (clinical cure + mycological eradication), n (%)	161 (62.1)	154 (59.0)
Failure, n (%)	74 (28.6)	82 (31.4)
Indeterminate/missing, n (%)	24 (9.3)	25 (9.6)
Risk difference (95% CI)	—	3.1% (−4.2, +10.5)

Rules: 

• Composite endpoint per SAP = clinical cure + mycological eradication. 

• Failure classification: death before end of therapy (EOT) or need for rescue antifungal = failure. 

• Indeterminate: no post-baseline culture and no documented clinical cure. 

Explanation and interpretation 

This table summarises the composite endpoint of global response at EOT, a standard secondary outcome in invasive candidiasis trials. 

• Success rates: 62.1% for Fungimycin vs 59.0% for caspofungin. 

• Failures: Slightly fewer with Fungimycin (28.6% vs 31.4%). 

• Indeterminate cases: Low and similar (9–10%), suggesting no systematic missing-data issue. 

• Risk difference: +3.1% (95% CI −4.2, +10.5), favouring Fungimycin but with CI crossing 0. 

This indicates Fungimycin achieved numerically higher global response, but the CI overlaps, so no statistical superiority is demonstrated. 

Writer’s checks 

• Confirm that the endpoint definition aligns with SAP and the protocol (i.e., “rescue = failure”, “death = failure”). 

• Verify handling of indeterminate cases is consistent. 

• Note whether this secondary endpoint is alpha-controlled; if not, interpret as descriptive only. 

• Ensure subgroup response rates (e.g. by species, neutropenia) do not reveal clinically relevant imbalances. 

CSR narrative example 

“At the end of therapy, 62.1% of Fungimycin patients achieved global response compared with 59.0% of caspofungin patients. Failure occurred in 28.6% vs 31.4%, and indeterminate outcomes were reported in ~9% of patients in both groups. The risk difference of 3.1% (95% CI −4.2, +10.5) favoured Fungimycin but did not demonstrate superiority. The composite endpoint followed prespecified SAP rules, with deaths and rescue antifungal use counted as failures.” 

Table 14.3.1 — Key Secondary Endpoint: Time to Blood-Culture Clearance (mITT Population) 

Metric	Fungimycin	Caspofungin
Median time (days), KM (95% CI)	4.2 (3.8, 4.8)	5.1 (4.5, 5.9)
Cox HR (95% CI)	1.15 (0.95, 1.38)	Reference
Log-rank p-value	0.15	—

Censoring rules: 

• If no clearance, censor at last culture date. 

• Clearance = first of two consecutive negative blood cultures collected ≥24 hours apart. 

Explanation and interpretation 

This table evaluates time to clearance of Candida from the bloodstream, a microbiological endpoint of interest in invasive candidiasis. 

• Median clearance time: Fungimycin = 4.2 days vs caspofungin = 5.1 days. This suggests slightly faster clearance with Fungimycin. 

• Hazard ratio: 1.15 (95% CI 0.95, 1.38). This favours Fungimycin (HR >1 = faster clearance), but CI includes 1.0, so not statistically significant. 

• Log-rank test: p = 0.15, supporting that the observed difference may be due to chance. 

Overall, Fungimycin shows a numerical trend towards faster clearance, but not statistically significant. 

Writer’s checks 

• Confirm censoring definition matches SAP (e.g. two negatives ≥24 h apart, last culture date for censored). 

• Verify the hierarchical testing plan: Was this endpoint alpha-controlled, or was it descriptive only? 

• Ensure narrative wording reflects non-significance “numerical trend” rather than “superior”. 

• Check for consistency with Kaplan–Meier plots (figures often accompany this table). 

CSR narrative example 

“Median time to blood-culture clearance was 4.2 days (95% CI 3.8–4.8) in the Fungimycin arm and 5.1 days (95% CI 4.5–5.9) in the caspofungin arm. The Cox hazard ratio was 1.15 (95% CI 0.95–1.38), favouring Fungimycin but not statistically significant (log-rank p=0.15). These results suggest a numerical trend towards faster clearance with Fungimycin, though superiority was not demonstrated.” 

Table 15.1.1 — Safety Overview (All Treated; Safety Set) 

Safety Parameter	Fungimycin (N=258)	Caspofungin (N=260)
Any TEAE, n (%)	163 (63.2)	163 (62.7)
Grade ≥3 TEAE, n (%)	49 (18.9)	51 (19.6)
Any SAE, n (%)	25 (9.7)	26 (10.0)
Discontinued due to TEAE, n (%)	11 (4.2)	10 (3.8)
Deaths on study, n (%)	5 (1.9)	6 (2.3)
AESI: Hepatic events, n (%)	12 (4.7)	8 (3.1)

Safety window: From the first dose until 30 days after the last dose. 

Relatedness: Based on investigator assessment. 

AESIs (Adverse Events of Special Interest): Pre-specified = hepatic events and QT prolongation. 

Explanation and interpretation 

This table provides a high-level safety overview, summarising adverse events, serious events, discontinuations, deaths, and AESIs. 

• Any TEAE: Incidence almost identical between Fungimycin (63.2%) and caspofungin (62.7%). Suggests overall tolerability was similar. 

• Grade ≥3 TEAEs: Comparable across arms (18.9% vs 19.6%). No major imbalance in severe toxicity. 

• SAEs: Occurred in ~10% of patients in both groups, balanced across arms. 

• Discontinuations due to TEAEs: Low (4–5%), with no meaningful difference. 

• Deaths: Rare (≤2.5%) and balanced (5 vs 6). Likely disease-related rather than drug-related. 

• AESIs (hepatic events): Slightly higher in Fungimycin (4.7% vs 3.1%). While not a major imbalance, this warrants cross-checking against laboratory shift tables (ALT/AST elevations) and individual narratives to confirm clinical relevance. 

Overall, the data suggest comparable safety profiles, with a minor signal for hepatic events in Fungimycin requiring further exploration. 

Writer’s checks 

• Confirm denominators align with Safety Set (≥1 dose). 

• Verify safety window (first dose to +30 days) matches protocol and SAP. 

• Ensure that relatedness assessments are consistent with the investigator’s attributions. 

• Cross-reference AESI (hepatic events) with lab shift/ECG tables for triangulation. 

• Note if further tables (e.g. TEAEs by SOC/PT, lab abnormalities) support or refute observed imbalances. 

CSR narrative example 

“In the Safety Set, overall TEAE incidence was similar between Fungimycin (63.2%) and caspofungin (62.7%). Grade ≥3 TEAEs were reported in 18.9% and 19.6% of patients, respectively. SAEs occurred in ~10% of patients in both groups, and discontinuations due to TEAEs were infrequent (≤5%). On-study deaths were balanced (1.9% vs 2.3%), reflecting underlying disease severity. Hepatic events, a pre-specified AESI, were slightly more frequent with Fungimycin (4.7% vs 3.1%), a finding to be further contextualised with laboratory shift data.” 

Table 15.2.4 — Most Common TEAEs (≥5% in any arm) by Preferred Term (Safety Set) 

Preferred Term (MedDRA)	Fungimycin (N=258)	Caspofungin (N=260)
Diarrhoea	41 (15.9)	34 (13.1)
ALT increased	32 (12.4)	27 (10.4)
Nausea	29 (11.2)	28 (10.8)
Rash	25 (9.7)	22 (8.5)
Headache	23 (8.9)	21 (8.1)
Anaemia	19 (7.4)	22 (8.5)
Pyrexia	18 (7.0)	20 (7.7)
AST increased	15 (5.8)	12 (4.6)

Explanation and interpretation 

This table lists the most common treatment-emergent adverse events (TEAEs) occurring in ≥5% of patients in either arm. 

• Gastrointestinal events (diarrhoea, nausea) were the most frequent, affecting ~13–16% of patients across both arms, and were balanced. 

• Skin and systemic events (rash, pyrexia, headache) were common but comparable between arms (~7–10%). 

• Anaemia occurred at similar rates (7.4% vs 8.5%). 

• Hepatic laboratory abnormalities (ALT/AST increases) appeared slightly more frequent in Fungimycin patients (12.4%/5.8%) compared with caspofungin (10.4%/4.6%). 

Overall, the TEAE profile is broadly consistent across treatment groups, with no clear imbalances except for a numerical trend towards higher hepatic enzyme elevations in Fungimycin. 

Writer’s checks 

• Cross-check against lab shift tables (Table 15.3.2) to see if enzyme increases reached ≥3×ULN. 

• Confirm whether hepatic TEAEs led to discontinuations (Table 15.1.1). 

• Align counts with AESI hepatic events reported earlier (Table 15.1.1). 

• Ensure consistent MedDRA coding and treatment windows. 

CSR narrative example 

“The most frequent TEAEs (≥5% incidence) were diarrhoea, nausea, rash, headache, anaemia, and pyrexia, with similar rates across treatment groups. Hepatic laboratory abnormalities (ALT and AST increases) were reported slightly more often in Fungimycin patients (12.4% and 5.8%) compared with caspofungin (10.4% and 4.6%), although these were generally of low grade. The overall pattern of common TEAEs was consistent between groups.” 

Table 15.3.2 — Laboratory Shifts: Hepatic Enzymes (Worst-on-Treatment; Safety Set) 

Parameter	Shift Category	Fungimycin (N=258)	Caspofungin (N=260)
ALT	<ULN → ≥3×ULN	6 (2.3)	4 (1.5)
ALT	≥ULN & <3×ULN → ≥3×ULN	3 (1.2)	2 (0.8)
AST	<ULN → ≥3×ULN	5 (1.9)	3 (1.2)
Bilirubin	<ULN → ≥2×ULN	1 (0.4)	1 (0.4)
Hy’s-Law candidates (ALT/AST ≥3×ULN + bilirubin ≥2×ULN; ALP not elevated)	—	0	0

Explanation and interpretation 

• ALT elevations: Slightly more frequent with Fungimycin (2.3% + 1.2%) vs caspofungin (1.5% + 0.8%). 

• AST elevations: Observed in 1.9% vs 1.2% — again, slightly higher with Fungimycin. 

• Bilirubin increases: Rare and identical across groups (0.4% each). 

• Hy’s Law candidates: None in either arm → reassuring for absence of drug-induced liver injury (DILI). 

Overall: Both drugs show a low frequency of hepatic lab shifts, with numerically more ALT/AST elevations in Fungimycin but no Hy’s Law signals. 

Writer’s checks 

• Confirm that ALT/AST increases correspond to TEAE “ALT increased” and “AST increased” in Table 15.2.4. 

• Verify AESI hepatic events in Table 15.1.1 align with these lab findings. 

• Note whether any patients with ALT/AST ≥3×ULN discontinued (check discontinuation TFLs). 

• Cross-reference SAP for lab grading thresholds to ensure correct categorisation. 

CSR narrative example 

“Laboratory shifts in hepatic enzymes were infrequent. ALT elevations to ≥3×ULN occurred in 2.3% of Fungimycin-treated patients and 1.5% of caspofungin-treated patients; an additional 1.2% and 0.8% of patients, respectively, had pre-existing mild elevations that increased to ≥3×ULN. AST elevations were reported in 1.9% of Fungimycin patients and 1.2% of caspofungin patients. Bilirubin increased to ≥2×ULN in one patient in each group. No Hy’s Law cases were identified. These findings are consistent with TEAE reports of ALT and AST increases and with the predefined AESI hepatic events.” 

Table 16.1 — Microbiology Outcomes (Micro Set) 

Outcome	Fungimycin (n/N, %)	Caspofungin (n/N, %)
Eradication at EOT (overall)	171/233 (73.4)	165/231 (71.4)
C. albicans eradication	94/112 (83.9)	92/110 (83.6)
C. glabrata eradication	43/62 (69.4)	42/64 (65.6)
C. parapsilosis eradication	21/34 (61.8)	21/36 (58.3)
C. tropicalis eradication	12/21 (57.1)	11/20 (55.0)
C. auris eradication	3/6 (50.0)	3/6 (50.0)

Explanation and interpretation 

• Overall eradication rates: Comparable between arms (73.4% Fungimycin vs 71.4% caspofungin). Suggests both drugs had similar microbiological efficacy at end of therapy (EOT). 

• Species-specific outcomes: 

• C. albicans: Highest eradication (~84% in both arms), reflecting expected susceptibility. 

• C. glabrata: Moderate eradication (~69–70%), again balanced between arms. 

• C. parapsilosis and C. tropicalis: Lower eradication (~58–62%), consistent with their known reduced susceptibility to echinocandins/novel antifungals. 

• C. auris: Lowest eradication (50% in both arms), but numbers are very small (n=6 per arm), limiting interpretation. 

Overall pattern: No clinically meaningful imbalances between treatment arms. Results reflect typical variability by Candida species. 

Writer’s checks 

• Confirm denominators (n/N) match central lab culture-confirmed baseline isolates. 

• Highlight C. auris as numerically weaker but avoid over-interpretation due to small sample size. 

• Ensure definitions of eradication (per protocol/SAP) are clearly stated in CSR methods (e.g. “two consecutive negative cultures” or “absence at EOT”). 

• Align with clinical endpoints: if failure due to persistent infection overlaps with “lack of efficacy” discontinuations, mention cross-link. 

CSR narrative example 

“Microbiological eradication rates at end of therapy were similar between Fungimycin (73.4%) and caspofungin (71.4%). Species-specific outcomes showed the highest eradication for C. albicans (83.9% vs 83.6%) and the lowest for C. auris (50.0% in both groups), although the latter was based on only six patients per arm. Eradication rates for C. glabrata, C. parapsilosis, and C. tropicalis were lower than for C. albicans but balanced across treatment arms. These results indicate comparable microbiological efficacy of the two treatments, with no clinically meaningful imbalances by species.” 

Table 7.2 — Protocol Deviations (All Randomised) 

Deviation Category (major)	Fungimycin (N=259)	Caspofungin (N=261)
Inclusion criteria not met (major)	3 (1.2)	4 (1.5)
Prohibited concomitant antifungal prior to first dose	2 (0.8)	3 (1.1)
Endpoint window violation (primary)	1 (0.4)	1 (0.4)
Total major deviations	9 (3.5)	10 (3.8)

Explanation and interpretation 

• Overall incidence: Major protocol deviations were rare (3.5% vs 3.8%), showing good compliance with trial procedures. 

• Types of deviations: 

• Inclusion criteria not met: 1–2% of patients, balanced between arms. 

• Prohibited concomitant antifungal use: very few cases (≤1%). 

• Primary endpoint window violations: only 1 case per arm. 

• Balance across arms: No meaningful imbalance, reducing risk of bias. 

• Impact on analysis sets: These deviations may have led to exclusions from the Per-Protocol (PP) set, but because rates are low and balanced, the overall efficacy conclusions are unlikely to be affected. 

Writer’s checks 

• Confirm that PP set definitions (exclusions for major deviations) are consistent with the SAP. 

• Cross-check which patients were excluded from PP analyses and whether exclusions differ by treatment arm. 

• Note whether any endpoint-related violations (e.g. visit outside time window) affect sensitivity analyses for the primary endpoint. 

CSR narrative example 

“Major protocol deviations were infrequent and occurred at comparable rates in the Fungimycin (3.5%) and caspofungin (3.8%) groups. The most common deviations were failure to meet inclusion criteria (1.2% vs. 1.5%) and the use of prohibited concomitant antifungals prior to the first dose (0.8% vs. 1.1%). Only one patient in each arm had a primary endpoint window violation. Overall, the low frequency and balanced distribution of deviations indicate minimal impact on study validity or the robustness of efficacy analyses.” 

Listing extracts (examples) 

• L-SAE: Patient-level SAE narratives (onset, seriousness, action taken, outcome). 

• L-Death: Death narratives with cause, timing, relation, and source records. 

• L-ProtDev: Detailed major/minor protocol deviations. 

• L-Micro: Per-patient species/MIC/eradication trajectory. 

Writer’s checks: Use listings to validate anomalies seen in tables and to support narrative clarity. 

How to turn TFLs into regulator-ready text (step-by-step) 

Use a repeatable micro-pattern for every endpoint: S.O.S. = Subject → Outcome → Statistic/Decision. 

Primary endpoint (CSR; ICH E3) 

Text: 

“In the mITT population, Day-30 all-cause mortality occurred in 18/259 (6.9%) patients with Fungimycin and 22/261 (8.4%) with caspofungin. The risk difference was −1.5% (95% CI −5.5, +2.5); the upper CI bound was below the pre-specified 10% non-inferiority margin, thereby demonstrating non-inferiority. Kaplan–Meier estimates were consistent (7.0% vs 8.3%).” 

(Trace to Table 14.1.1; SAP §8.2.1 for imputation.) 

Key secondary (CSR) 

Global response at EOT: 

“62.1% vs 59.0%; RD 3.1% (95% CI −4.2, +10.5); composite rules applied (rescue = failure).” 

(Trace to Table 14.2.1; SAP composite rules.) 

Time to culture clearance: 

“Median 4.2 vs 5.1 days; HR 1.15 (95% CI 0.95, 1.38); log-rank p=0.15.” 

(Trace to Table 14.3.1; censoring rules per SAP.) 

Safety (CSR) 

“TEAEs occurred in 63.2% vs 62.7%; Grade ≥3 in 18.9% vs 19.6%; SAEs in 9.7% vs 10.0%. Discontinuations due to TEAEs were 4.2% vs 3.8%. Hepatic AESIs were 4.7% vs 3.1%; no Hy’s-Law cases were identified. Lab shift analysis showed ALT ≥3×ULN in 2.3% vs 1.5% and AST ≥3×ULN in 1.9% vs 1.2%, generally transient.” 

(Trace to Tables 15.1.1, 15.2.4, 15.3.2.) 

CTD Module 2.7.3 / 2.7.4 (concise) 

• 2.7.3 (Efficacy): “Fungimycin was non-inferior to caspofungin for Day-30 mortality (6.9% vs 8.4%; RD −1.5%, 95% CI −5.5, +2.5). Global response at EOT was 62.1% vs 59.0% (RD 3.1%; 95% CI −4.2, +10.5). Median time to blood-culture clearance 4.2 vs 5.1 days (HR 1.15; 95% CI 0.95, 1.38).” 

• 2.7.4 (Safety): “The overall safety profile was broadly comparable between arms. No Hy’s-Law cases were identified. Hepatic laboratory increases were infrequent and predominantly transient.” 

Briefing document seed (example) 

Q1 (Primary NI Margin): “Does the CHMP agree that the 10% NI margin remains acceptable given the observed RD −1.5% (95% CI −5.5, +2.5) in mITT?” 

Position: “Yes; NI criterion met. Sensitivity analyses support the conclusion.” 

Quality control checklist (before you release drafts) 

• Traceability: Every number in text maps to a TFL cell; all set/time-window labels present. 

• Estimand alignment: Intercurrent events handled per SAP (treatment-policy vs composite); clearly stated. 

• Multiplicity: Superiority comments only if allowed by hierarchy; otherwise “numeric trend”. 

• Safety consistency: Same denominators and windows across safety tables; AESIs align with lab/ECG findings. 

• Microbiology clarity: Species-level statements neutral; eradication definitions present. 

• Version control: TFL version matches your draft version; changes logged. 

Common pitfalls (and how to avoid them) 

• Numbers drift between CSR and 2.7.x → lock a single “source of truth” spreadsheet and a Traceability Matrix. 

• Endpoint definition creep → paste exact SAP wording into your internal notes; never paraphrase loosely. 

• Over-claiming beyond alpha control → use “numerically higher/lower” or “trend” for exploratory analyses. 

• Mismatched denominators/windows → label everything (set + window) in text and tables. 

• Untriangulated safety signals → cross-check TEAEs, AESIs, labs, and discontinuations together. 

Conclusion 

TFLs are the evidence backbone of the CSR and CTD summaries. Your job is to read them like a regulator: verify analysis sets and rules, respect the SAP, and produce neutral, precise, and fully traceable narratives. Starting from subject disposition and baseline balance through primary/secondary efficacy, safety, and microbiology, the steps above show how to turn a voluminous TFL package into a coherent, inspection-ready story that supports benefit–risk assessment. 

If you hold to three principles—accuracy, clarity, consistency—you’ll deliver documents that clinical, stats, PV, and regulatory colleagues can sign off swiftly, and that agencies can navigate with confidence. 

References  

Donnelly, J.P., Chen, S.C., Kauffman, C.A., Steinbach, W.J., Baddley, J.W., Verweij, P.E., Clancy, C.J., Wingard, J.R., Lockhart, S.R., Groll, A.H. and Sorrell, T.C., 2020. Revision and update of the consensus definitions of invasive fungal disease from the European Organisation for Research and Treatment of Cancer and the Mycoses Study Group Education and Research Consortium. Clinical Infectious Diseases, 71(6), pp.1367–1376. https://doi.org/10.1093/cid/ciz1008 

International Council for Harmonisation (ICH), 1995. E3: Structure and Content of Clinical Study Reports. Available at: https://database.ich.org/sites/default/files/E3_Guideline.pdf [Accessed 10 September 2025]. 

International Council for Harmonisation (ICH), 2016. M4E(R2): The Common Technical Document — Efficacy: Structure and Content of Clinical Summary/Overview. Available at: https://database.ich.org/sites/default/files/M4E_R2__Guideline.pdf [Accessed 10 September 2025]. 

International Council for Harmonisation (ICH), 1998. E9: Statistical Principles for Clinical Trials. Available at: https://database.ich.org/sites/default/files/E9_Guideline.pdf [Accessed 10 September 2025]. 

International Council for Harmonisation (ICH), 2019. E9(R1): Addendum on Estimands and Sensitivity Analysis in Clinical Trials. Available at: https://database.ich.org/sites/default/files/E9-R1_Step4_Guideline_2019_1203.pdf [Accessed 10 September 2025]. 

Disclaimer 

Fungimycin, data, and TFLs in this post are fictional and provided for educational purposes to illustrate regulatory writing practice. They do not represent real clinical results, products, or submissions.