For eyes below 22mm, Hoffer Q has a strong evidence base. Modern formulas including Barrett and Kane are competitive. Nanophthalmos is a separate category.
Short eyes present a disproportionate challenge in IOL power calculation. Altered ocular geometry, increased sensitivity to effective lens position prediction errors, and the higher IOL powers required all combine to make refractive outcomes less predictable than in eyes of average axial length. Formula selection is one modifiable variable — but the evidence is more nuanced than a simple formula recommendation.
The refractive consequence of axial length measurement error increases as eyes become shorter. When a high-power IOL is required, small biometric errors translate into larger dioptric consequences than in average eyes. Effective lens position prediction is also more sensitive to anterior chamber depth assumptions in eyes with altered biometric geometry.
Short eyes are not a uniform group. Eyes in the 21–22mm range behave differently from eyes below 21mm, and nanophthalmic eyes below 20mm represent a separate clinical category with substantially greater anatomical complexity, crowded anterior segments, elevated angle closure risk, and significantly worse refractive predictability. Formula performance data from the 21–22mm range should not be extrapolated directly to extreme short eyes.
Hoffer Q was developed specifically to improve prediction in shorter eyes, using a personalized anterior chamber depth model rather than a purely regression-derived constant. Eom et al. (2014, JCRS) provided contemporary comparative validation in eyes below 22mm, finding Hoffer Q outperformed SRK/T and Holladay 1 in this range, with the advantage most pronounced in eyes below 21mm. Aristodemou et al. (2011, JCRS) analyzed outcomes across 8,108 eyes and similarly found Hoffer Q performed better than SRK/T and Holladay 1 in shorter eyes.
Legacy effective lens position prediction models may become less reliable at biometric extremes. Holladay 1 may be less predictable than Hoffer Q or newer formulas in very short eyes, particularly below 21mm. With careful constant optimization from a surgeon's own outcomes data, performance can be improved — but without personalization, the population-derived starting point is less accurate in extreme biometry.
Modern formulas including Barrett Universal II, Kane, and Olsen incorporate additional biometric variables — such as lens thickness and more sophisticated ELP modeling — that improve performance across the axial length range including short eyes. Kane et al. (2016, JCRS) found newer formulas often comparable or superior to Hoffer Q in short eyes, though results varied by cohort. Melles et al. (2018) similarly found Barrett Universal II among the strongest performers across axial length extremes.
The Olsen formula deserves specific mention. Its C-constant approach to ELP prediction uses the preoperative ACD and lens thickness to estimate postoperative lens position and has shown strong performance in biometrically challenging eyes. Surgeons working with extreme short eyes should consider Olsen or Barrett Universal II alongside Hoffer Q as part of a multi-formula approach.
Axial length measurement — optical biometry is strongly preferred in short eyes. Constant optimization — personalized constants from the surgeon's own outcomes data reduce systematic bias and matter more in short eyes. Target refraction — some surgeons intentionally target slight myopia in short eyes as a buffer against hyperopic surprise. Patient counseling — residual refractive error is somewhat more likely in short eyes even with optimal planning.
IOLDx Clinical displays SRK/T, Hoffer Q, Holladay 1, and Haigis side by side for every lens.
Open IOL Planner →