Evidence Paper Review

 Quick Explanation Copied

 Long Explanation

Study Design and Key Findings

Between February 1999 and September 2002, researchers at the Hereditary Cancer Institute of Riga Stradins University collected cancer family histories from 865 colorectal cancer (CRC) patients (~30% of all new Latvian CRC cases) using a structured questionnaire covering first- and second-degree relatives. Families were classified against Amsterdam II, late-onset HNPCC, suspected HNPCC, familial colorectal cancer (FCC), and cancer familial aggregation (CFA) criteria, then underwent DHPLC screening plus sequencing of MLH1, MSH2, and MSH6 heteroduplexes, with MSH2/MSH6 immunohistochemistry (IHC) on 27 available tumor samples Methods of the Latvian hereditary CRC program.

Reported results: 463 patients (53.53%) reported at least one malignancy in a blood relative; only 3 pedigrees (0.35%) fulfilled Amsterdam II criteria (without histopathologic confirmation); 14 (1.62%) met suspected HNPCC criteria; 41 (4.74%) matched FCC; and 69 (8%) showed CFA. Molecular testing yielded four germline mutations: MLH1 c.1409+1G/A (exon 12), MSH2 c.508C>T (Gln170Stop, exon 3), MSH2 c.1786delAAT (596delAsn, exon 12), and MSH6 c.1815-1816delTA (exon 4) Classification results and detected mutations.

The Age 30–39 Signal

The most actionable finding is the mutation-detection gradient by age: 2 of 11 probands aged 30–39 (18.2%) carried a germline mutation, versus 0.3–0.4% in older age bands and zero below age 30. The authors accordingly propose targeted screening of all CRC patients aged 30–39 Age-stratified mutation detection rates. However, this is built on just 2 mutations in 11 patients—an extremely small denominator—and the 0-mutation finding in the <30 group rests on only 2 cases, so the "age 30–39 priority" is hypothesis-generating, not established.

Critical Appraisal

Strengths: A genuinely first-of-its-kind national effort for Latvia, ethically approved, with written informed consent, an international collaboration (Pomeranian Medical University, Poland), and consistent with the Finnish prospective estimate of 0.7–2.4% HNPCC frequency Cross-country comparison of HNPCC frequency.

Blind spots the authors partially acknowledge: pedigree reliability is explicitly flagged as poor—many families had few relatives and sparse medical documentation, and none of the three Amsterdam II pedigrees could be histopathologically verified, which formally disqualifies them by their own criteria. The cohort is hospital-based and non-consecutive (patient refusals, limited research capacity), so representativeness is asserted by demographic comparability rather than demonstrated. Sequencing was restricted to DHPLC heteroduplexes, missing variants DHPLC is known to miss; large rearrangements in MSH2/MLH1 were not tested—a documented cause of under-detection. Only 27 IHC analyses were performed out of ~865 cases. The claim of "one of the lowest reported HNPCC incidences" may therefore reflect detection-method limitations as much as biology. No data-availability statement or public deposit exists, and no confounder-adjusted analysis was performed.

Missing information: microsatellite instability status, PMS2 testing, tumor sequencing of IHC-negative cases beyond one MSH6 mutation, follow-up outcomes of counseled relatives, and per-patient consent/uptake rates.

Confidence: Moderate. The descriptive counts are clearly reported, but the ~2% incidence estimate carries wide, unquantified uncertainty (authors themselves state 0.5–2%), and the age-stratification recommendation rests on n=2 mutations. The conclusion would change if a modern, comprehensively sequenced, randomly sampled Latvian CRC cohort (with MMR rearrangement testing and verified pathology) found a substantially different incidence or mutation spectrum.

 BGPT Paper Review

Study Novelty

First population-scale hereditary CRC registry for Latvia (novel geography and founder-mutation context), but the clinical and molecular framework (Amsterdam II criteria, MLH1/MSH2 sequencing) was already standard by 2003.

Scientific Quality

Reasonable methods (consent, ethics approval, DHPLC+sequencing, IHC) but three Amsterdam II pedigrees lacked pathological verification—disqualifying them by their own criteria; DHPLC misses variants; no MSI testing; no PMS2; non-consecutive hospital sampling; incidence reported as a wide 0.5–2% range without confidence intervals.

Study Generality

Findings are nation-specific; the ~2% incidence and Latvian mutation spectrum add regional epidemiological value but do not generalize beyond the Baltic context, and only 4 mutations prevent spectrum conclusions.

Study Usefulness

Practically useful for Latvian genetic-service planning and the age 30–39 screening hypothesis, but the actionable claim rests on 2 of 11 mutations and needs modern, full-spectrum validation before guideline adoption.

Study Reproducibility

Methods (questionnaire content, criteria, DHPLC/sequencing, IHC) are described at moderate detail, but no data deposit, no per-family raw data, and no analytic code limit independent reproduction.

Explanatory Depth