INTRODUCTION

Renal osteodystrophy is a condition affecting patients with a chronically impaired renal function.1–3 In these patients, regular dialysis treatment can only partly compensate for the lost physiological function, and several biochemical and hormonal modifications occur. Among them, hyperphosphatemia, hypocalcemia, and low plasma levels of calcitriol play a key role in conditioning renal osteodystrophy.1–3 High bone turnover, due to elevated parathormone (PTH) plasma levels (a condition known as secondary hyperparathyroidism), and osteomalacia, due to a mineralization defect, can often be present. These modifications in bone metabolism provoke morphological changes in the skeleton, leading even to macroscopic bone deformities.1–3

Osteodystrophy involves numerous body districts, and it can also be observed in the craniofacial skeleton. Several reports underlined the relevant clinical signs, radiological appearances, and pathological features.1–4 One of the extreme manifestations of this condition involves the progressive hypertrophy of the craniofacial bones, leading to a form of leontiasis ossea.2 Literature reports on uremic leontiasis ossea are scanty, but maxillary and mandibular hypertrophy with increased palatal dimensions, a flattening of the nasal bridge together with a widening of the nares, and prominent cheek bones have all been described in these patients.2 Recently, Schmidt et al3 reported that neither the incidence of uremic leontiasis ossea nor subclinical craniofacial hypertrophy in uremic patients is established. The modifications of the facial skeleton will involve also its soft tissue covering, with an alteration in the appearance of the patient.

Current technology provides several noninvasive image analysis systems for indirect computerized facial anthropometry including stereophotogrammetry, laser scanning, range cameras, optoelectronic instruments, and electromagnetic digitizers.5–12 In addition, ultrasonography allows facial anthropometry to be performed even during intrauterine life.1314 These instruments provide the three-dimensional coordinates of selected landmarks, and euclidean geometric calculations can be used to obtain three-dimensional linear distances of selected facial structures, as well as facial areas and volumes.5–101215

The quantitative assessment of craniofacial variations is commonly used for the characterization of diseased subjects.679–111617 To the best of our knowledge, no previous study quantitatively assessed the soft tissue facial structure of uremic patients without or with leontiasis ossea.

In the present study, the facial soft tissues of a group of patients with chronic renal insufficiency with regular dialysis treatment were measured in three-dimensional space with a computerized digitizer, and their facial dimensions (linear distances, ratios, angles, volumes) calculated and compared with a reference normal population by using z-scores.

MATERIALS AND METHODS

Patients

Data on 20 patients with chronic renal insufficiency (10 males, 10 females; aged 53–81 years) were collected in the present study (Table 1). All patients were Caucasian (northern Italians) and were under treatment at the Dialysis Unit of San Paolo Hospital (Milan). All were in a clinical stable condition during at least the last two years before evaluation, and no patient had received craniofacial surgical procedures. Patients were selected for enrollment according to the degree of hyperparathyroidism. Ten of them had serum intact parathormone (iPTH) lower than 300 pg/mL (mean 175.4, SD 96.0), whereas the remaining had serum iPTH higher than 500 pg/mL (mean 651.5, SD 142.9) (mean of the last one year of observation). Plasma iPTH levels were measured with the IRMA assay (Nichols Institute Diagnostic, San Juan Capistrano, Calif), which has a reference range of 15–65 pg/mL. Normal parathyroid status, for the uremic population, was considered for PTH <300 pg/mL. Hyperparathyroidism status was considered for PTH values >500 pg/mL.

TABLE 1.  Analyzed Patients^a

View Fullscreen

All facial measurements were collected on the patients during the day after a dialysis procedure. Reference data were collected on 34 normal subjects (12 males, 22 females) of the same ethnic group and age. The normal subjects were either staff or relatives of the laboratory staff. No subjects with a previous history of craniofacial trauma or congenital anomalies were included in the reference group. All the analyzed individuals gave their informed consent to the experiment.

Collection of three-dimensional facial landmarks

The data collection procedure was done in two separate steps and was followed by off-line calculations.8918 Initially, a single experienced operator located a set of 50 soft tissue landmarks for each subject by inspection or palpation (or both) and marked them on the cutaneous surface using an eyeliner. The subjects sat relaxed in a position suitable for a correct identification of facial features. A total of 50 soft tissue landmarks were collected, which are identified in Figure 1.18 Midlandmarks were also mathematically derived as the midpoint between two homologous landmarks, and noted as landmark_m.

View Figure

• Download as Powerpoint
• Download Figure

In the second step, the three-dimensional coordinates of the facial landmarks were obtained with a computerized electromagnetic digitizer (3Draw, Polhemus Inc, Colchester, Vt). During data collection, the subjects sat in a natural head position in a wooden chair, with the head fixed to the chair back by a wooden and foam headframe. They were asked to close their eyes, to keep their teeth in contact (centric occlusion), and to not move for the duration of the data collection.8 Using the instrument stylus, a single operator digitized the marked landmarks according to a standardized sequence devised to reduce data collection time while the subjects sat motionless. Data collection took approximately one minute. The files of the three-dimensional (x, y, z) coordinates were obtained and stored on magnetic media. The subject was dismissed. Subsequently, computer programs devised and written by one of the authors were used for all the subsequent off-line calculations.

The reproducibility of landmark identification, marker positioning, and data collection procedure were previously reported and found to be reliable, with Dahlberg's errors on 50 landmarks of 1.20 mm (men) and 0.95 mm (women), corresponding to 1.04% and 1.05% of the relevant nasion-midtragion distances (n-t_m).8

RESULTS

In the present group of uremic patients, all the mean z-scores of the linear distances (except lower face height) were larger than in the reference subjects (Table 2). The z-scores were significantly different from zero in half the measurements. Overall, facial height (nose, mandibular ramus), width (skull base, nose), and depth (midface) increased 0.44–1.25 times in comparison with healthy, normal individuals of comparable age and of the same sex.

TABLE 2.  Descriptive Statistics of z-scores Computed in the Ure mic Patients

View Fullscreen

All facial volumes were significantly larger in the patients than in the reference subjects, and the largest difference was found for the maxillary volume. The ratio of mandibular to maxillary volume was significantly reduced in all patient groups.

Five of the measured facial angles tended to be reduced (more acute) in the uremic patients, whereas the mandibular corpus convexity in the horizontal plane was increased. On no occasion were the z-scores significantly different from zero.

Some significant linear correlations were found between patient characteristics and the anthropometric variables (Table 3). In particular, a longer duration of renal insufficiency was related to a longer facial height, a larger maxillary volume, a more different mandible-to-maxilla volume ratio, and a differently shaped mandible (convexity in the horizontal plane, gonial angles). A similar trend was found for nose height, where the duration of renal insufficiency explained about 19% of its variability. A longer history of regular dialysis treatment had nearly the same correlations.

TABLE 3.  Significant Linear Correlation Analyses Between Patient Characteristics and Anthropometric Variables

View Fullscreen

Higher iPTH levels significantly correlated with a longer nose, an increased facial lower third, and a larger right side gonial angle. In addition, calendar age was related to face height (an inverse relationship), mandibular corpus length, and both gonial angles.

DISCUSSION

Anthropometry gives an objective aid to the qualitative appraisal of soft tissue anatomy of the head and face, supplying the clinician with useful indications about the anatomical structures and regions that differ the most from the norm.679–111617

In the present study, a three-dimensional, noninvasive system allowed a fast (in less than one minute of data collection, one subject's complete set of data are available off-line in 10 minutes), low-cost, and quantitative analysis of the soft tissue facial characteristics in a group of adult uremic patients on regular dialysis treatment. Indeed, soft tissues can be studied with noninvasive methods, and the assessments can be performed frequently without any additional biological burden to the patients.8

Notwithstanding the substantial skeletal modifications found in patients with an impaired renal function,1–3 which can lead to massive modifications even in facial appearance,2 quantitative assessments of the overlying soft tissue morphology have been sparingly performed. In particular, until now, no detailed analyses of the variations in facial dimensions and shape in adult uremic patients have been published.

The soft tissue data collected on the present group of patients were compared with the values obtained in normal subjects of the same age, sex, and ethnic group by calculating z-scores. The use of z-scores allows to standardize data collected on individuals of different sex and age and to deal with a single group of patients in the assessment of statistical significances.917

Overall, the chronic renal insufficiency patients had larger, wider, and deeper faces than their normal peers. Although the standardized differences were only partly significant when single distances (height, width, depth) were considered, the analysis of global values (volumes) reached the statistical significance on all occasions.

Probably, the three-dimensional assessment of volume magnified the tendency of larger facial dimensions, where the interindividual variability prevented the assessment of a statistical significance. The increment in facial dimensions was particularly evident in the facial middle third (maxilla), leading to a significant inversion of the mandibular-to-maxillary ratio. The larger maxilla was accompanied by a tendency to more prominent lips (reduced interlabial angle). Some of these facial modifications (facial and nose height, maxillary dimensions, mandibular shape) were significantly related to the clinical characteristics of the patients (duration of renal insufficiency, duration of dialysis, iPTH levels).

The present data are difficult to compare with previous reports because literature references on the modifications of facial structures during chronic renal insufficiency deal mainly with the radiological appearance,1–4 and the extreme manifestation of leontiasis ossea seems to have been described in only nine patients.23 Cutaneous alterations, a common problem in uremic patients,19 have been described only on a pathological base.

Overall, the key macroscopic features of leontiasis ossea, as described by Lee et al2 and found in the patients in the present study, include a maxillary and mandibular skeletal hypertrophy as well as prominent cheek bones, and these features are significantly related to the soft tissue appearance. Increased palatal dimensions may explain the more prominent lips. The widening of the nostrils is directly measured by the significantly increased soft tissue nose width.

Lee et al2 reported that sex, race, age, length of dialysis treatment, or of the renal insufficiency did not seem to be predictive factors for the development of leontiasis ossea. In contrast, we did not assess patients with a fully developed alteration of the craniofacial skeleton but subjects who may be at risk of developing this further complication of their principal disease. In the patients in the present study, significant correlations were found between the anthropometric variables and the duration of the clinical history (renal insufficiency and dialysis), as well as the anthropometric variables and the mean levels of iPTH. These relationships were found in both the maxilla and the mandible (chin and gonial angle). The effect of race could not be assessed because all patients were of the same ethnic group. Age and sex were removed from the analysis by using the z-scores. Nevertheless, between 50 and 80 years of age some significant modification in mandibular morphology was found. These age-related trends were in the same direction as those apparently provoked by the clinical history of the patients and should be analyzed in detail in a future investigation.

Among the limitation of the present study, it has to be mentioned that facial volumes were obtained from a geometrical model of face where only some discrete landmarks have been sampled, and the surfaces between contiguous landmarks that are actually curved have been approximated by linear planes. Therefore, the calculations cannot provide the actual anatomical measurements but only an approximation of their values.15 Nevertheless, the same approximation was used for both the patients and the reference subjects, and only the difference between the two values was considered. Furthermore, the method has been used in our laboratory for approximately a decade with valuable results.15

In addition, the patients in the study represent a convenience sample, and the results must be considered with caution. Only further investigations performed on larger groups of patients may allow a deeper insight into the facial appearance of uremic subjects.

In conclusion, the method used in the current investigation allowed a simple, low-cost, fast, and noninvasive examination of the patients and provided a quantitative assessment of the standardized deviation from the norm to be used also on a longitudinal base. We believe that the method could be profitably used to assess the modifications of the facial structures in uremic patients even before the development of manifest clinical signs of renal osteodystrophy, thus allowing a finer monitoring of the additional burdens experienced by them.