Chronic Inhibition of cGMP Phosphodiesterase 5A Improves Diabetic Cardiomyopathy

The researchers tested whether a daily low-dose regimen of sildenafil (a PDE5 inhibitor, commonly known as Viagra) could reverse or slow the early structural and functional heart changes seen in diabetic cardiomyopathy — a condition where the heart muscle stiffens and pumps less efficiently due to diabetes, even in the absence of blocked arteries.

**Intervention:** 100 mg sildenafil per day, split into three doses (25 mg morning, 25 mg afternoon, 50 mg at bedtime) for 3 months.

**Comparator:** Identical-looking placebo capsules taken on the same schedule.

**Primary outcomes:** Changes in left ventricular (LV) torsion (the twisting motion of the heart during contraction, measured in degrees) and LV strain (the percentage of shortening of heart muscle fibers during contraction).

**Secondary outcomes:** LV mass-to-volume ratio (a measure of concentric hypertrophy — thickening of the heart walls), N-terminal pro–B-type natriuretic peptide (NT-proBNP, a blood marker of heart strain), vascular endothelial growth factor (VEGF), monocyte chemotactic protein-1 (MCP-1, an inflammatory marker), transforming growth factor-beta (TGF-β, a fibrosis marker), blood pressure, endothelial function (flow-mediated dilation), and metabolic parameters (HbA1c, glucose, lipids).

**Sample size:** 59 men (30 sildenafil, 29 placebo) who completed the study.

**Population:** Men aged 35–75 years (mean 60.3 ± 7.4 years) with type 2 diabetes for at least 1 year (mean duration 6.5 years), recruited from outpatient clinics at Policlinico Umberto I, Sapienza University Hospital of Rome between 2008 and 2009.

**Inclusion criteria:** Type 2 diabetes ≥1 year; HbA1c ≤10%; normal blood pressure or treated hypertension with target <130/80 mmHg; body mass index <40.

**Exclusion criteria:** Use of exogenous insulin, thiazolidinediones (a class of diabetes drugs), or spironolactone; prior or current use of PDE5 inhibitors; substance abuse; history of cardiovascular disease, proliferative retinopathy, or autonomic neuropathy; symptoms or signs of ischemic heart disease on stress echocardiography or cardiac MRI at enrollment; contraindications to sildenafil or MRI.

**Key baseline characteristics:** Mean HbA1c 7.6–7.8%; mean BMI 28; mean ejection fraction 53–55% (normal range); all had evidence of diabetic cardiomyopathy on cardiac MRI (reduced circumferential strain < -15%, increased LV torsion > 14°, increased LV mass-to-volume ratio > 1.5 g/mL).

**Setting:** Single academic medical center in Rome, Italy.

**Cardiac magnetic resonance imaging (CMR):** Performed on a 1.5-Tesla Siemens Avanto scanner with ECG triggering. Two types of sequences were used:

1. **Cine MRI (steady-state free precession):** 10–12 short-axis slices covering the entire left ventricle from base to apex. Used to measure LV volumes (end-diastolic volume, end-systolic volume, stroke volume), ejection fraction, LV mass, and chamber geometry (septal thickness, transverse diameter). All volumes indexed to body surface area.

2. **Myocardial tagging (CMR-tagged imaging):** A grid pattern was "stamped" onto the heart muscle at end-diastole using radiofrequency saturation pulses. Images were acquired every 40 milliseconds thereafter. By tracking how the grid lines deformed during contraction, the researchers calculated:

- **Circumferential strain (ε, %):** The percentage shortening of heart muscle fibers around the circumference of the ventricle. More negative values = better contraction (normal is approximately -18 to -20%).

- **LV torsion (θ, degrees):** The net twist of the apex relative to the base during systole. Measured by comparing rotation at the base versus apex. Normal is approximately 10–14°; higher values indicate compensatory hyper-twisting.

3. **Late gadolinium enhancement:** After IV injection of 0.2 mmol/kg gadolinium-BOPTA, T1-weighted images were taken 10–15 minutes later to detect areas of myocardial fibrosis (scar tissue), defined as signal intensity >5 standard deviations above normal myocardium.

**Blood markers:** NT-proBNP, VEGF, MCP-1, TGF-β, HbA1c, glucose, lipids, homeostasis model assessment (HOMA) index for insulin resistance.

**Vascular function:** Flow-mediated dilation (FMD) of the brachial artery — a measure of endothelial function.

**Blood pressure:** 24-hour ambulatory blood pressure monitoring.

**Adverse events and compliance:** Monthly follow-up visits, weekly telephone interviews, monthly pill counts.

**Study design:** Randomized, double-blind, placebo-controlled, parallel-group clinical trial.

**Randomization:** Permuted-block randomization with 1:1 allocation using randomly sized blocks of 4–8 patients. The allocation list was generated by dedicated software. Randomization was not stratified by any baseline variables.

**Blinding:** Triple-blinding — patients, outcome assessors (radiologists analyzing CMR images, biologists running blood assays), and care providers were all blinded. Allocation codes were kept in sequentially numbered sealed envelopes by a pharmacist not involved in the study. Codes were broken only 3 months after the last patient completed the study. Blinding was maintained for outcome assessors until full statistical analysis was completed.

**Duration:** 3 months of treatment, with 1 month of baseline screening and 1 month of post-treatment follow-up (total study duration 5 months per patient).

**Dosing rationale:** 100 mg/day split into three doses (25 mg + 25 mg + 50 mg) to maintain relatively stable plasma levels throughout the day, rather than the single 50–100 mg dose used for erectile dysfunction. The bedtime dose was higher to cover the overnight period when cardiac remodeling processes may be most active.

**Statistical approach:** Primary analysis compared change from baseline to 3 months between sildenafil and placebo groups using analysis of covariance (ANCOVA) or Mann-Whitney U test for non-normally distributed variables. Results reported as mean ± SD change scores. No correction for multiple comparisons was mentioned. Sample size calculation was not explicitly stated, but 59 completers is modest for detecting small effect sizes.

**What this design can prove:**

Causal relationship between sildenafil and cardiac changes (randomized, placebo-controlled)

Specificity of effect to PDE5 inhibition (double-blind eliminates expectation bias)

Independence from blood pressure or metabolic changes (measured and found unchanged)

Direct myocardial action (since afterload, endothelial function, and metabolism didn't change)

**What this design cannot prove:**

Mechanism of action (can't distinguish between direct cardiomyocyte effects vs. indirect effects through coronary microcirculation or other unmeasured pathways)

Long-term outcomes (3 months is short; no data on whether effects persist, reverse, or translate to reduced heart failure hospitalizations or mortality)

Generalizability to women (all male), younger patients, or those with more advanced heart disease

Dose-response relationship (only one dose tested)

Whether effects are specific to diabetic cardiomyopathy or apply to other forms of heart disease

Durability after stopping treatment (no washout or follow-up beyond 1 month)

**Major methodological strengths:**

Randomized, double-blind, placebo-controlled design

Objective, quantitative primary outcomes (CMR with tagging) rather than subjective symptoms

Comprehensive measurement of potential confounders (blood pressure, endothelial function, metabolism)

Industry-independent funding and conduct

Strict inclusion/exclusion criteria to isolate diabetic cardiomyopathy from ischemic or hypertensive heart disease

**Major methodological weaknesses:**

Small sample size (59 completers) — risk of type II error for secondary outcomes

Single-center study — limited generalizability

No correction for multiple comparisons — some secondary findings may be false positives

No a priori sample size calculation reported

All-male sample — cannot generalize to women

Short duration (3 months) — no data on long-term safety or efficacy

No mechanistic tissue analysis (no biopsies) — mechanism inferred but not proven

Potential for unmeasured confounders despite randomization (groups well-matched at baseline, but small samples can hide imbalances)

**Primary outcomes (cardiac kinetics):**

**LV torsion change:** Sildenafil group decreased by -3.89 ± 3.11° (from ~18.4° to ~14.5° — toward normal); placebo group increased by +2.13 ± 2.35° (worsening). Between-group difference: ~6.0° (P < 0.001).

**Circumferential strain change:** Sildenafil group improved by -3.30 ± 1.86% (more negative = better contraction, from ~-12.6% to ~-15.9%); placebo group worsened by +1.22 ± 1.84% (less negative = worse contraction). Between-group difference: ~4.5% (P < 0.001).

**Secondary outcomes (cardiac structure and function):**

**LV mass-to-volume ratio:** Improved by 6.5 ± 11% in sildenafil vs. placebo (P = 0.021). This indicates regression of concentric hypertrophy — the heart walls became less thick relative to chamber size.

**LV end-diastolic volume:** Increased by ~8 mL in sildenafil vs. placebo (P = 0.04), suggesting improved filling capacity.

**LV mass:** Decreased by ~6 g in sildenafil vs. placebo (P = 0.03), indicating regression of hypertrophy.

**Ejection fraction:** No significant change (both groups remained in normal range ~55–60%).

**Blood markers:**

**MCP-1 (inflammatory marker):** Decreased by -75.30 ± 159.28 pg/mL in sildenafil vs. placebo (P = 0.032).

**TGF-β (fibrosis marker):** Decreased by -5.26 ± 9.67 ng/mL in sildenafil vs. placebo (P = 0.009).

**NT-proBNP (heart strain marker):** Trend toward decrease but not statistically significant (P = 0.08).

**VEGF:** No significant change.

**Vascular and metabolic outcomes (all non-significant):**

**Flow-mediated dilation:** No change — endothelial function unaffected.

**24-hour blood pressure:** No change — afterload unaffected.

**HbA1c, glucose, insulin, HOMA index:** No change — metabolic control unaffected.

**Lipids:** No change.

**Safety and tolerability:**

No serious adverse events reported.

Mild adverse events (headache, flushing, dyspepsia) occurred in ~20% of sildenafil group vs. ~10% of placebo — all transient and self-limiting.

No dropouts due to adverse events.

**In plain English:**

**Torsion (heart twist):** The diabetic heart compensates for poor contraction by twisting more — like wringing out a wet towel harder because the fibers are stiff. Sildenafil reduced this compensatory twisting by about 4 degrees over 3 months, bringing it closer to normal (~14° vs. ~18° at baseline). This is roughly a 20% improvement — moving from "clearly abnormal" toward "borderline normal."

**Strain (heart shortening):** Normal heart muscle shortens by about 18–20% during contraction. These diabetic patients started at ~12.6% shortening. After sildenafil, they improved to ~15.9% — a 26% relative improvement in contractile function. This is roughly equivalent to recovering about one-third of the lost function.

**Mass-to-volume ratio:** This dropped by 6.5% — meaning the heart walls became less thick relative to the chamber size. In practical terms, the heart was becoming less "stiff" and more compliant, able to fill with blood more easily.

**To put it in perspective:** These improvements are comparable to what you might see after 6–12 months of intensive exercise training or significant weight loss in early heart disease. The effect size is moderate — not dramatic, but clinically meaningful for a drug that works through a completely different mechanism than standard heart failure medications.

**Importantly:** These changes happened without any improvement in blood pressure, blood sugar, cholesterol, or endothelial function. The drug appeared to work directly on the heart muscle itself, not through secondary effects on the blood vessels or metabolism.

**Author-acknowledged limitations:**

Small sample size (59 patients) — limits statistical power for secondary outcomes and subgroup analyses.

Short duration (3 months) — cannot assess long-term effects on clinical outcomes like heart failure hospitalization or mortality.

All-male sample — results may not apply to women with diabetic cardiomyopathy.

Single-center study — may not generalize to other populations or settings.

No myocardial biopsies — mechanism of action inferred from imaging and blood markers, not directly demonstrated in human heart tissue.

No dose-response data — only one dose (100 mg/day) tested.

**Critical reader observations:**

**No correction for multiple comparisons:** With multiple primary and secondary outcomes, some significant findings (especially the weaker ones like MCP-1 at P = 0.032 and TGF-β at P = 0.009) could be false positives. The primary outcomes (torsion and strain) are robust at P < 0.001, but the secondary markers should be interpreted cautiously.

**No sample size calculation:** The study may be underpowered for some outcomes. The non-significant trend for NT-proBNP (P = 0.08) might become significant with a larger sample — or might not.

**Selection bias:** Patients were highly selected — no insulin use, no thiazolidinediones, no cardiovascular disease, normal ejection fraction. This limits generalizability to the broader diabetic population, many of whom have more advanced disease or are on insulin.

**Compliance monitoring:** Pill counts are unreliable — patients can discard pills. No blood levels of sildenafil were measured to confirm adherence.

**Industry funding not disclosed:** The authors state the study was conducted without industry support, but sildenafil is a commercial product (Pfizer). No conflict of interest statement was provided in the abstract.

**No intention-to-treat analysis:** The results are reported for completers only (59 of presumably more randomized). If dropouts differed between groups, this could bias results.

**Short follow-up after treatment:** Only 1 month of post-treatment follow-up — no data on whether effects persist or reverse after stopping.

**No assessment of diastolic function directly:** While torsion and strain reflect both systolic and diastolic function, the study did not report standard diastolic parameters like E/A ratio or deceleration time from echocardiography.

For someone running their own n=1 experiment (with medical supervision — PDE5 inhibitors have contraindications and interactions):

### What to test

**Intervention:** Daily low-dose sildenafil (or another PDE5 inhibitor like tadalafil). The study used 100 mg/day split into three doses. For self-experimentation, a reasonable starting protocol would be 25–50 mg once daily (tadalafil 5–10 mg daily has a longer half-life and may be more convenient).

**Comparator:** Before-and-after measurements (n=1 design) or alternating periods of drug vs. no drug (if medically appropriate).

### Minimum meaningful duration

**3 months minimum.** The study showed significant changes at 3 months. Shorter durations (weeks) are unlikely to produce measurable cardiac remodeling. Some effects may appear earlier (4–6 weeks), but the full antiremodeling effect likely requires 8–12 weeks.

### What to measure

**Primary metric:** Cardiac function — ideally via echocardiography with speckle-tracking strain imaging (available at most cardiology clinics). Key parameters:

- Global longitudinal strain (GLS) — the most reproducible measure of myocardial deformation. Normal is approximately -18 to -22