Quality Celestara

AI-Powered Annual Product Quality Review

Annual Product Quality Review is the single largest recurring compliance documentation exercise in pharmaceutical quality—and it is overwhelmingly manual at most companies. QA teams spend months compiling data from disconnected systems, assembling spreadsheets, and producing reports that are already stale by the time they are complete. Celestara transforms APQR from a reactive annual burden into continuous product quality monitoring, with AI agents that aggregate data in real time, detect adverse trends as they emerge, and generate audit-ready reports on demand.

Key Highlights

01 Reduce APQR preparation from weeks to days
02 Continuous trend detection across all products
03 Automated data aggregation from 6–8 systems
04 Audit-ready reports with statistical analysis

Regulatory Context

What FDA Inspectors Cite

21 CFR 211.180(e) Incomplete Review
1 / 6

Annual product review deficient — fails to define batches, complaints, recalls, or investigations for the review period

The firm’s annual product reviews did not define batches manufactured during the year being reviewed. There was no documented evidence of rejected batches, complaint investigations, recalls, returned batches, or manufacturing/laboratory investigations. The APR lacked specificity and focused only on an assessment of applicable procedures rather than actual product quality data.

Brands International Corporation · 2024-06-21
21 CFR 211.180(e) Missing Reviews
2 / 6

No annual product review records maintained for aseptically produced drug products

The firm does not maintain written records for annual product reviews conducted for aseptically produced drug products. Over 14 product formulations—including Amiodarone, Heparin, Vancomycin, and Norepinephrine—had no APR documentation, leaving no evidence that quality standards were evaluated annually.

LEESAR, Inc. · 2025-05-09
21 CFR 211.166 Stability Program
3 / 6

No stability testing program designed to assess stability characteristics of drug products

The Quality Unit failed to ensure drug products sold into the US market are stable throughout their labelled shelf life. No testing program was established to assess stability characteristics, meaning critical trending data for APQR—long-term and accelerated stability results, adverse trends, and shelf-life projections—simply did not exist.

Torrent Pharmaceuticals Limited · 2024-06-12
21 CFR 211.22 Data Integrity
4 / 6

Environmental monitoring data entered into uncontrolled spreadsheets with inaccurate trending reports

Environmental monitoring data was manually entered from collection sheets into uncontrolled electronic spreadsheets used to generate reports and graphical analyses. Multiple discrepancies were identified between source data and reports. Trend analysis summary reports contained inaccurate mold recovery data, excluded relevant deviations, and had mislabeled graphs—making CAPA effectiveness assessment impossible.

Dr. Reddy’s Laboratories, LTD, Biologics · 2025-09-12
21 CFR 211.103 Batch Yield Tracking
5 / 6

Quality Unit failed to calculate and document yield percentages for packaging batches

The firm failed to calculate and document yield percentages for approximately hundreds of packaging batches destined for the US market. Percent yields were not calculated, documented, or reviewed for any potential issues with the specific and separate packaging operation of each batch—data that is a core component of any meaningful annual product review.

Mylan Laboratories Limited · 2024-06-26
21 CFR 211.166 Stability Testing
6 / 6

Written stability program not followed — failed to test stability samples for OTC products

During stability data review of several OTC products, the firm failed to test stability samples at required intervals. The written stability testing program was not followed, meaning APQR had no reliable stability data to evaluate whether products remained within specification throughout their shelf life.

Sato Pharmaceutical Co., Ltd. · 2026-01-05

The Problem

Why APQR Takes Months Instead of Days

Challenge 1 1 / 7

Critical Data Is Unstructured and Inaccessible

  • Batch records exist as scanned PDFs, handwritten logbooks, and free-text deviation narratives
  • Complaint descriptions, investigation conclusions, and CAPA rationales are unstructured text buried in documents
  • Standard reporting tools require structured database fields—they cannot extract insights from PDFs, emails, or narrative records
Challenge 2 2 / 7

Data Fragmented Across 6–8 Systems

  • Batch records in MES, analytical results in LIMS, deviations in QMS, stability in a separate system
  • Manual CSV exports and copy-paste into Excel is the industry-standard approach
  • No single source of truth—data reconciliation alone can take weeks per product
Challenge 3 3 / 7

4–6 Weeks Per Product, Per Year

  • A company with 50–100 products can spend 2,000–5,000+ person-hours annually on APQR
  • Coordination required across QA, QC, Validation, Production, Engineering, and Stability teams
  • The Q1 quality crunch diverts resources from production support and deviation investigations
Challenge 4 4 / 7

Data Compilation Without Critical Analysis

  • FDA expects meaningful analysis—not just data dumps—yet most firms lack statistical tools
  • Adverse trends in yield, impurities, or dissolution go undetected in spreadsheet compilations
  • No cross-product correlation: shared equipment or raw material issues are missed
Challenge 5 5 / 7

Insights Arrive 3–9 Months Late

  • A process drift in March may not trigger a CAPA until January of the following year
  • Annual cadence means problems persist for months before corrective action is initiated
  • Continuous process verification (FDA Stage 3) requires real-time trending, not annual snapshots
Challenge 6 6 / 7

Multi-Site Format Inconsistency

  • Cross-site comparison is impossible when data formats and statistical methods differ
  • Products manufactured at one site but tested at another require manual data consolidation
  • EU PQR, US PAR, and ICH Q7 require different data elements—multiplying the documentation burden
Challenge 7 7 / 7

Spreadsheet-Based Data Integrity Risks

  • Manual data entry from source systems creates undetectable transcription errors
  • No version control—drafts circulate across departments with no audit trail
  • Qualification of Excel spreadsheets as GxP tools adds validation burden without reducing risk

Annual Product Quality Review in Celestara

1

Continuous Data Aggregation Across All Systems

Celestara continuously pulls batch yields, analytical results, deviations, CAPAs, complaints, stability data, and environmental monitoring from all connected systems. When APQR time arrives, the data is already there—normalized, validated, and complete.

Cross-System Extraction Format Normalization Completeness Validation
2

Automated Statistical Trending and Analysis

AI generates control charts, calculates Cpk/Ppk indices, and performs trend analysis across all critical quality attributes for every batch—not at the end of the year, but continuously. Adverse trends trigger alerts the moment they emerge.

Cpk Calculation Control Chart Generation Trend Detection Anomaly Flagging
3

Stability and Shelf-Life Intelligence

The platform monitors all ongoing stability studies, flags out-of-trend results before they become out-of-specification, and projects shelf-life trajectories. Every stability data point is automatically linked to its product APQR.

Shelf-Life Projection OOT Detection Adverse Trend Flagging
4

Cross-Product and Cross-Site Pattern Detection

Celestara correlates quality data across products that share equipment, raw materials, or processes. When a yield drift affects multiple products on the same line, the system connects the dots—turning isolated observations into systemic insights.

Shared Equipment Analysis Raw Material Correlation Multi-Site Trending
5

Audit-Ready Report Generation

When the review period ends, Celestara generates a complete APQR draft with all required data elements, statistical visualizations, narrative summaries, and regulatory-specific formatting—ready for QA review and approval, not weeks of manual assembly.

Template Compliance Narrative Drafting Regulatory Formatting

The Solution

How Leucine Solves This

Purpose-built tools that address both the FDA compliance gaps and the operational barriers to effective annual product quality review.

Celestara Quality Review

Structured and Unstructured Data, Unified Automatically

Celestara continuously aggregates data from all connected systems—and unlike standard reporting tools, its AI agents can extract insights from scanned batch records, free-text deviation narratives, handwritten logbooks, and PDF reports. When the review period ends, every data element is already collected, normalized, and validated.

  • AI extraction from unstructured sources: scanned PDFs, handwritten logs, free-text narratives
  • Automated ingestion from MES, LIMS, ERP, QMS, and stability systems
  • Real-time data completeness monitoring with gap alerts
AI Statistical Intelligence

Meaningful Trend Analysis That Goes Beyond Data Tables

AI generates control charts, calculates process capability indices (Cpk/Ppk), and performs continuous trend detection across all critical quality attributes. Adverse trends are flagged the moment they emerge—not months later during annual compilation.

  • Automated Cpk/Ppk calculations and control chart generation for every CQA
  • Cross-product correlation for shared equipment, raw materials, and processes
  • Anomaly detection that identifies subtle drifts before they become OOS events
MES

Standardized Reviews Across Every Manufacturing Site

Deploy consistent APQR templates, statistical methodologies, and approval workflows across all facilities. Every site follows the same review process, making cross-site trending meaningful and regulatory submissions consistent.

  • Standardized APQR templates aligned to FDA PAR and EU PQR requirements
  • Multi-site data aggregation with harmonized trending methodologies
  • Automated report generation with regulatory-specific formatting
FDA Tracker

Know What Inspectors Expect Before They Arrive

Monitor how FDA cites APQR deficiencies across the industry. Understand which review gaps draw 483s—from missing stability data to inadequate statistical trending—so your team knows exactly what inspectors evaluate first.

  • Real-time tracking of 483 observations related to 211.180(e) and annual reviews
  • Warning letter analysis for APQR and stability program deficiencies
  • Benchmarking your review practices against industry enforcement trends

Results

Measurable Impact

Real results from organizations using Celestara for annual product quality review

85%
APQR Preparation Time
Reduction in time to prepare annual product quality reviews, from weeks to days per product
99%
Data Completeness
Automated verification that all required data elements are collected before report generation
12x
Trend Detection
Faster identification of adverse trends through continuous monitoring versus annual review
3,000+
QA Hours Recovered
Person-hours per year redirected from manual APQR compilation to proactive quality improvement

Next Step

Get Started

Stop spending months compiling annual reviews from spreadsheets. Celestara continuously aggregates quality data from all your systems, performs real-time statistical trending, and generates audit-ready APQR reports on demand—transforming a reactive compliance exercise into continuous product quality intelligence.

Get Started
Exit